Quillaia extracts (synonyms, soapbark extracts, Quillay bark extracts, bois de Panama, Panama bark extracts, quillai extracts) are obtained by aqueous extraction of the milled inner bark or of the wood of pruned stems and branches of Quillaja saponaria Molina (family Rosaceae). The term ‘quillaia’ refers to the dried inner bark of the tree, which is a large evergreen with shiny, leathery leaves and a thick bark, native to China and several South American countries, principally Bolivia, Chile, and Peru.

Unpurified extracts contain over 60 triterpenoid saponins, consisting predomi-nantly of glycosides of quillaic acid. Polyphenols and tannins are major components. Some simple sugars and calcium oxalate are also present. The saponin concentration of freshly prepared, unpurified extracts is 190–200 g/kg of solids (about 20%). The extracts are treated with ‘stabilizing agents’ such as egg albumin and polyvinyl-pyrrolidone and then filtered through diatomaceous earth. The stabilizing agents remove substances that would probably precipitate during storage, such as protein–polyphenol complexes. After filtration, the liquid is concentrated, and the concentrate may be sold as such (solids constituting about 550 g/L) or be spray-dried and sold as a powder containing carriers such as lactose and maltodextrin. The unpurified extracts are used in food applications, primarily for their foaming properties.

Semi-purified powdered extracts are produced by subjecting unpurified extracts to ultra-filtration or affinity chromatography to remove most non-saponin solids, such as polyphenols. These semi-purified extracts have higher saponin concentrations (750–800 g/kg of solids; about 80%) and better emulsifying properties than unpurified extracts.

Highly purified extracts are produced for use as adjuvants in the production of animal and human vaccines and not for food use. These products generally contain more than 90% saponins.

In previous evaluations, the Committee considered data that related to unpurified quillaia extracts. Quillaia extract was reviewed toxicologically by the Committee at its twenty-sixth meeting (Annex 1, reference 59). The available toxicological data included adequate lifetime studies in mice and rats, from which a NOEL was identified. However, in the absence of data, no specifications were prepared, and, hence, no ADI could be allocated. At its twenty-ninth meeting (Annex 1, reference 70), the Committee prepared new tentative specifications and established an ADI of 0–5 mg/kg bw. The Codex Committee on Food Additives and Contaminants (CCFAC) (2000) at its thirty-second session requested the Expert Committee to re-evaluate all relevant information on the toxicity and, in particular, the intake of quillaia extracts. No new data were submitted to the Committee. Published reports on quillaia extracts or specific saponins that provided information relevant to a toxicological assessment of quillaia extracts were evaluated at the present meeting.

Quillaia extracts are mixtures of biologically active compounds, including saponins, tannins, polyphenols, and calcium oxalate. The saponins present in quillaia extract have a variety of biological activities: haemolytic, cytotoxic, immune-enhancing (adjuvant), mucosal irritation and inflammation, and anti-hypercholesterolaemic. The biological activities and the potency of individual saponins vary widely and depend in vivo primarily on the route of administration.

Saponins extracted with ethanol from soapbark trees and administered orally to rabbits with experimental atherosclerosis resulted in an increased ratio of plasma lecithin to cholesterol, normalized blood cholesterol levels, and decreased elevated blood pressure. Subcutaneous injection of the saponin extract did not affect the atherosclerotic symptoms (Efimova et al., 1966).

Groups of 12 Swiss mice (strain and sex unstated) were injected into the footpad with aliquots of seven different 5% extracts of Q. saponica bark. Animals that lived longer than 24 h and controls were killed. The degree of oedema in the mice, haemolysis of rabbit erythrocytes, and the adjuvant strength of the various extracts in stimulating immunity to staphylococcal toxin were measured with the same quillaia extracts. No relationship was found between these parameters (Richou et al., 1965).

The effect of a range of saponins, including crude quillaia saponin on gut permeability was assessed by monitoring the steady-state glucose transfer potential in vitro in sections of jejunal mucosa from male Wistar rats. The individual saponins elicited widely different responses in the small gut and these were significantly affected by pH, concentration, chemical formula, and the presence of other materials in the solution. Quillaia extract caused a reduction in transmural potential difference comparable to that observed with the basic glycoalkaloids in potato and tomato and the complex bisdesmosides from Gypsophilia and alfalfa. These saponins were all more potent than the saponins from soya, which showed only weak activity. The reduction in transmural potential difference has been associated with increased uptake of both passively permeable sugars and large compounds and with a loss of the ability of the mucosa to accumulate actively transported organic species(Gee et al., 1989).

When ‘pure’ saponin from Q. saponaria was administered subcutaneously to mice that had been immunized with bovine serum albumin or Crotalus durissus (South American rattlesnake) venom, a significantly higher antibody titre against the antigens was found in the sera of immunized mice than in animals receiving the antigens alone. Tumour necrosis factor activity was increased in mice immunized with bovine serum albumin and/or saponin, but interferon-gamma was produced only in mice immunized with both bovine serum albumin and saponin (Gebara et al., 1995).

The effect of the Quillaia saponin fractions QH-A, QH-B, and QH-C and a crude Quillaia saponin extract (Spikoside) on haemolytic activity, cytotoxicity, and macromolecular synthesis was studied in vitro. A concentration of 5 µg/ml of QH-B or QH-C caused haemolysis of chicken erythrocytes after 1 h of incubation at 37 ^oC. QH-A was haemolytic at an approximately 10-fold higher concentration, 50 µg/ml. The crude extract caused haemolysis at a concentration of 20 µg/ml. No haemolytic activity was observed at concentrations <100 µg/ml when these preparations were incorporated into an immunostimulating complex matrix. Cytotoxicity was assessed by measuring intracellular dehydrogenase activity by a colorimetric method. Seeded WEHI 164 cells clone 13⁹ were incubated with various concentrations of the extracts for 2 h before analysis. QH-B and QH-C inhibited enzyme activity at a concentration of approximately 10 µg/ml and the crude extract at a concentration of approximately 20 µg/ml. QH-A was tolerated at concentrations <100 µg/ml. When QH-A, QH-B, QH-C, and the crude extract were incorporated into the immunostimulating complex matrix, the cells tolerated approximately 10-fold higher concentrations. Macromolecular synthesis was assessed by measuring incorporation of [³H]leucine and [³H]uridine into protein and RNA, respectively, in WEHI 164 cultured cells. Treatment with QH-B or QH-C at concentrations <10 µg/ml for 30 min had no affect on protein or RNA synthesis (Rönnberg et al., 1995).

Saponins vary widely in the kind and intensity of their biological activity. Some of the more important effects included haemolysis (strong in vitro, much weaker in vivo), local irritation, inflammation (intestine), anti-inflammatory and antimicrobial activity, cytotoxicity, and anti-hypercholesterolaemic in laboratory animals. Severe toxic effects reported after large oral doses were liver damage, respiratory failure, gastric pain, diarrhoea, haemolysis of erythrocytes, convulsions, and coma (Leung, 1980).

2.2.1 Acute toxicity

In mice, saponins extracted from the soapbark tree were less acutely toxic when administered orally (LD50, 1600 mg/kg bw) than when administered subcutaneously (650 mg/kg bw), intraperitoneally (280 mg/kg bw), or intravenously (280 mg/kg bw) (Efimova et al., 1966).

The acute toxicity of a Q. saponaria extract and of QS-7, QS-18, and QS-21 saponins isolated from the extract was investigated in groups of five CD-1 mice (sex not stated), 8–10 weeks of age given an intradermal concentration of 125, 250, or 500 µg of each saponin or the quillaia extract and monitored for 72 h. QS-18 was the most lethal of the substances tested, with 4/5, 5/5, and 5/5 deaths at the three doses, respectively. The authors stated that deaths occurred at a dose of QS-18 as low as 25 µg. QS-7 was not lethal at doses up to 500 µg, and QS-21 was lethal only at 500 µg, at which 1/5 mice died. The quillaia extract resulted in the deaths of 1/5, 2/5, and 4/5 mice at the three doses, respectively (Kensil et al., 1991).

The acute toxicity of subcutaneously administered QH-A, QH-B, QH-C, and crude Quillaia saponin (Spikoside) was tested in groups of 10 female ICR mice at doses of 50–400 µg. All mice survived after the highest dose of QH-A and QH-C, with no visible signs of toxicity. All mice given 400 µg QH-B died within 12 h, while seven mice survived injection with 200 µg. When the crude extract was administered, doses of 200 and 400 µg resulted in 6 and 10 deaths, respectively. All mice survived the lower doses of QH-B or the crude extract (Rönnberg et al., 1995).

A purified, toxic Quillaia triterpenoid fraction with strong adjuvant activity, designated QH-B, was used to study whether modification of the carbohydrate moiety with sodium periodate would alter the toxicity without harming the adjuvant activity and the cholesterol-binding capacity. Groups of 10 ICR female mice received single increasing subcutaneous doses of modified QH-B and were observed for 3 weeks. Unmodified QH-B and QH-B modified by treatment with 2.5–10 mmol/L sodium periodate had similar LD50 values, ranging from 29 to 70 µg per mouse. Treatment of the fraction with 25 or 50 mmol/L sodium periodate increased the LD50 of QH-B to 197 and 132 µg, respectively. The difference in the toxicity of the periodate-modified QH-B may have been due to alterations in the structure of the sugars galactose and xylose (Rönnberg et al., 1997).

2.2.2 Short-term studies of toxicity

Rats

Groups of 15 male and 15 female weanling CFE rats were housed five per cage and fed diets containing 0 (control), 0.6, 2, or 4% quillaia extract for 13 weeks. Groups of five male and five female rats from the same lot of animals were fed diets containing 0 (control), 2.0%, or 4.0% quillaia extracts for 2 or 6 weeks. Body weights and food intake were measured at the beginning and weekly throughout the study. Urine was analysed during the final week of the study. At sacrifice, the absolute and relative organ weights were determined, and tissues and organs from the group given 4% were analysed histologically. Haematological and serum chemical parameters were tested in all groups.

No abnormalities of behaviour or condition were seen in the rats receiving quillaia extract. The body weights of those at the highest concentration (4%) were significantly lower than those of controls up to day 78 in males, but only for the first 2 weeks in females. Food and water consumption were reduced in animals of each sex at all dietary levels, but by the end of the study the weights of the treated rats did not differ significantly from those of the controls. Feeding quillaia saponin did not affect the results of haematological examinations, serum or urine analyses, renal concentrating ability, or urinary cell excretion. The relative liver weight was reduced in males given 2 or 4% quillaia extract, and the relative stomach weight was increased in animals of each sex at the same concentrations. No histopathological effects attributable to treatment were found. The NOEL was 0.6% in the diet, equivalent to 400 mg/kg bw per day (Gaunt et al., 1974).

In a 90-day study to assess the safety of a saponin extract from Thea sinensis L., quillaia saponin (approximately 8.5% sapogenin; not clear whether this product meets current specifications) was used for comparison. Although the Committee examined this study, it was considered irrelevant to the toxicological assessment of quillaia extracts as sufficient data were not available on the specifications of the test material and because the animals were given the compound by gavage (Kawaguchi et al., 1994).

2.2.3 Long-term studies of toxicity

Mice

Groups of 48 male and 48 female TO strain mice were fed diets containing 0, 0.1, 0.5, or 1.5% quillaia extract for 84 weeks. The mice were observed regularly for abnormal condition or behaviour, and some males were weighed at intervals. Haematological parameters were measured at weeks 24, 56, and 84. No compound-related effects were reported. At the highest dose, male mice showed decreased weight gain. Quillaia at these concentrations had no adverse effect on condition, behaviour, or death rate. A detailed autopsy and histopathological examination of tissues and organs at the end of the study showed no compound-related effects. No carcinogenic effects were seen. The slightly lower body-weight gain of mice at the highest dietary concentration and some changes in organ weights, albeit of doubtful significance, resulted in a NOEL of 0.5%, equivalent to 700 mg/kg bw per day (Phillips et al., 1979).

Rats

Groups of 48 male and 48 female Wistar-derived rats from a specific pathogen-free breeding colony were housed in groups of four and fed diets containing 0 (control), 0.3, 1, or 3% quillaia extract for 108 weeks. Haematological examinations were made at weeks 15, 25, 52, and 108, and urine was analysed at weeks 13, 24, and 78. At the end of the study, a complete autopsy was carried out, including histological examination of tissues and organs.

Male rats at the highest dietary concentration had lower body weights than controls throughout the experiment, the differences being statistically significant between 10 and 22 months. Females at the lowest dietary concentration had significantly higher body weights than controls during the first 6 months of the study. The lower body weights and reduced incidence of glomerulonephrosis in male rats fed 3% quillaia were considered to be due to reduced food consumption. Preference tests run before the start of the 2-year study showed that rats avoided the diet containing quillaia extract. Haematological and urinary parameters were within the normal ranges.

In general, the incidence of histopathological lesions was similar in treated and control animals, only those of fibrosis of the heart and dilatation of the glands of the gastric mucosa in females at the lowest dietary concentration being greater than those of controls. These effects were considered to be fortuitous, as there was no dose–response relationship and no similar occurrence in males. A variety of benign and malignant tumours was found. The incidences of haemangiomas and haemangio-sarcomas in the lymph nodes were similar in control and treated animals. The only tumour for which the incidence was statistically significantly different from that in controls was thyroid adenoma, which occurred more frequently in females fed 1% quillaia in the diet. This finding was not considered to be related to treatment, as the incidence did not increase with the dietary concentration of quillaia extract, and the total incidence in rats fed 1% quillaia extract was not statistically significantly different from the total control incidence (Drake et al., 1982).

Quillaia extracts may be used as a foaming agent in soft drinks, such as ginger beer, root beer, and cream soda, in cocktail mixes, and as an emulsifier in other foods, such as baked goods, candies, frozen dairy products, gelatine, and puddings. The major food use is in soft drinks. Quillaia extracts are currently proposed for use in the Codex draft General Standard for Food additives (GSFA) at 500 mg/kg in food group 14.1.4 ‘Water-based flavoured drinks’, including ‘sport’ or ‘electrolyte’ drinks and particulated drinks.

Data on intake of quillaia extracts were submitted to the Committee by Australia and New Zealand, the United Kingdom, and the USA.

The budget method was used to decide whether the intake of quillaia extracts should be assessed further. The calculations indicated that the theoretical maximum level of use of quillaia extracts is 100 mg/kg, assuming that it is used in only half the beverages in a food supply and that the ADI is 0–5 mg/kg bw. As this theoretical level is lower than the proposed level of 500 mg/kg in the draft GSFA, further assessment of intake is needed. The draft GSFA proposes use of quillaia extract in one food group only, 14.1.4 ‘Water-based flavoured drinks’.

As quillaia extracts are proposed for use in a single food group, a reverse budget method was used to indicate the maximum amount of the food group that can be consumed before the ADI is exceeded. In this case, up to 600 g/day of water-based flavoured drinks could be consumed if quillaia extracts were used at a concentration of 500 mg/kg, assuming an ADI of 0–5 mg/kg bw and an average body weight of 60 kg, while a child of 15 kg can drink only 150 g/day before exceeding the ADI.

If quillaia extracts were used at 100 mg/kg (the maximum manufacturers’ use levels are 95 mg/kg in the United Kingdom and 100 mg/kg in the USA), up to 3000 g/day of water-based flavoured drinks could be consumed before the ADI was exceeded, assuming an ADI of 0–5 mg/kg bw, an average body weight of 60 kg, and up to 750 g/day for a 15-kg child.

Poundage (disappearance) data were available from the USA, based on information reported to the National Academy of Sciences (1989): 38 600 pounds (17 500 kg) of quillaia extracts were reported to have been used in food applications in 1987. The intake of quillaia extracts per capita was calculated to be 0.0055 mg/kg bw per day (0.1% ADI), assuming a body weight of 60 kg and 60% response to the survey (raw poundage data divided by 0.6 to account for underreporting). The per capita intake at the 90th percentile was calculated by multiplying by a factor of 2, to give 0.011 mg/kg bw per day or 0.2% ADI (use of factor discussed in WHO, 1987).

Estimates of the intake of quillaia extracts based on individual dietary records from national surveys in Australia and New Zealand, the United Kingdom, and the USA were based on consumption of the whole water-based flavoured drinks category or those soft drinks likely to contain the additive. The estimates and the assumptions made in deriving the estimates are summarized in Table 1.

    See Also:
       Toxicological Abbreviations
       Quillaia extracts (WHO Food Additives Series 17)
       QUILLAIA EXTRACTS (JECFA Evaluation)