WHO FOOD ADDITIVES
Safety evaluation of certain food additives
Prepared by the Sixty-third meeting of the Joint FAO/WHO
Expert Committee on Food Additives (JECFA)
Geneva, 2006 6
Extract, point 4 and 5 of 28 pages:
After oral administration, steviol glycosides are poorly absorbed in experimental animals and in humans.
Intestinal microflora metabolize steviol glycosides to the aglycone, steviol, by successive hydrolytic removal of glucose units. Data reviewed by the Committee at its current and fifty-first meetings (Annex 1, reference 149) indicated that this process is similar in rats and humans. The hydrolysis of rebaudioside A to steviol was slower than that of stevioside. In humans treated orally with stevioside, small amounts of steviol were detected in the plasma, with considerable interindividual variability. The major route by which steviol is metabolized in humans in vivo appears to be via conjugation with glucuronide and/or sulfate. Studies with liver microsomal preparations indicated that steviol is also metabolized to a number of hydroxy and dihydroxy derivatives via CYP-dependent pathways.
Stevioside and/or steviol affected a variety of biochemical parameters in models in vitro, indicating possible mechanisms of antihypertensive and antiglycaemic effects that involve modulation of ion channels. High concentrations (e.g. 1 mmol/l) of stevioside were required to produce a maximal increase in insulin secretion, while steviol was effective at a concentration that was about three orders of magnitude lower. Stevioside also affected a variety of biochemical parameters in different animal species in vivo, mostly with parenteral administration; these studies were considered by the Committee to be of limited relevance to dietary exposure.
No new long-term studies of toxicity or carcinogenicity were available at the present meeting. At its fifth-first meeting, the Committee noted that oral administration of stevioside (purity, 95.6%) at a dietary concentration of 2.5%, equal to 970 and 1100 mg/kg bw per day in male and female rats, respectively, for 2 years was not associated with toxicity. Reduced body-weight gain and survival rate were observed with stevioside at a dietary concentration of 5%. In a new study, stevioside was found to inhibit the promotion of skin tumours by TPA in a model of skin carcinogenesis in mice.
The Committee reviewed new data on genotoxicity that, considered together with data reviewed by the Committee at its fifth-first meeting, allowed a number of conclusions to be drawn. Stevioside and rebaudioside A have not shown evidence of genotoxicity in vitro or in vivo. Steviol and some of its oxidative derivatives show clear evidence of genotoxicity in vitro, particularly in the presence of a metabolic activation system. However, studies of DNA damage and micronucleus formation in rats, mice and hamsters in vivo indicate that the genotoxicity of steviol is not expressed at doses of up to 8000 mg/kg bw.
One new study of developmental toxicity was available at the present meeting.
Adverse effects on the reproductive apparatus, which could be associated with impaired fertility, were observed in male rats given a crude extract of S. rebaudiana, at a dose corresponding to 1.34 g of dried leaves. However, at its fifth-first meeting, the Committee reviewed a number of studies of reproductive and developmental toxicity with stevioside (purity, 90% or 96.5%). Doses of up to 2500 mg/kg bw per day in hamsters and 3000 mg/kg bw per day in rats had no effect in studies of reproductive toxicity. No teratogenic or embryotoxic effects were observed in rats given stevioside at a dose of up to 1000 mg/kg bw per day by gavage. The Committee considered that the adverse reproductive effects associated with administration
of aqueous extracts of S. rebaudiana, noted at the present and fifty-first meeting, were unlikely to be caused by steviol glycosides.
Stevioside is being investigated as a potential treatment for hypertension and diabetes. Administration of stevioside at a dose of 750 or 1500 mg per day for 3–24 months resulted in decreased blood pressure in hypertensive patients, with no adverse effects. These studies, in a limited number of subjects, provided some reassurance that stevioside at a dose of up to 25 mg/kg bw per day (equivalent to 10 mg/kg bw per day expressed as steviol) for up to 2 years shows no evidence of significant adverse effects in these individuals. There is no information on the effects of repeated administration of stevioside on blood pressure in normotensive individuals. A small study in 12 patients with type-2 diabetes showed that a single dose of 1 g of stevioside reduced postprandial glucose concentrations and had no effect on blood pressure.
The Committee evaluated information on intake of steviol glycosides, submitted by Japan and China. Additional information was available from a report on S. rebaudiana Bertoni plants and leaves that was prepared for the European Commission by the Scientific Committee on Food. All the intake results are presented in terms of equivalents of steviol, based on a conversion of 40% from the steviol glycoside, stevioside (relative molecular mass: steviol, 318, stevioside, 805).
The Committee used the GEMS/Food database to prepare international estimates of intake of steviol glycosides (as steviol). It was assumed that steviol glycosides would replace all dietary sugars, at the lowest reported relative sweetness ratio for steviol glycosides and sucrose, 200 : 1. The intakes ranged from 1.3 mg/ kg bw per day (African diet) to 3.5 mg/kg bw per day (European diet).
The Committee evaluated estimates of per capita intake derived from disappearance (poundage) data supplied by Japan and China. The Committee also evaluated estimates of intake of steviol glycosides based on the replacement of all dietary sugars in the diets for Japan and the USA. These results are summarized in Table 5.
The Committee concluded that the replacement estimates were highly conservative and that intake of steviol glycosides (as steviol) would be likely to be 20–30% of these values.
The Committee noted that most of the data requested at its fifty-first meeting, e.g. data on the metabolism of stevioside in humans, and on the activity of steviol in suitable studies of genotoxicity in vivo, had been made available.
The Committee concluded that stevioside and rebaudioside A are not genotoxic in vitro or in vivo and that the genotoxicity of steviol and some of its oxidative derivatives in vitro is not expressed in vivo. The no-observed-effect level (NOEL) for stevioside was 970 mg/kg bw per day in a long-term study evaluated by the Committee at its fifty-first meeting.
The Committee noted that stevioside has shown some evidence of pharmacological effects in patients with hypertension or with type-2 diabetes at doses corresponding to about 12.5–25 mg/kg bw per day (equivalent to 5–10 mg/kg bw per day expressed as steviol). The evidence available at present was inadequate to assess whether these pharmacological effects would also occur at lower levels of dietary exposure, which could lead to adverse effects in some individuals (e.g. those with hypotension or diabetes). The Committee therefore decided to allocate a temporary acceptable daily intake (ADI), pending submission of further data on the pharmacological effects of steviol glycosides in humans.
A temporary ADI of 0–2 mg/kg bw was established for steviol glycosides, expressed as steviol, on the basis of the NOEL for stevioside of 970 mg/kg bw per day (or 383 mg/kg bw per day expressed as steviol) in the 2-year study in rats and a safety factor of 200. This safety factor incorporates a factor of 100 for inter- and intraspecies differences and an additional factor of 2 because of the need for further information. The Committee noted that this temporary ADI only applies to products complying with the specifications.
The Committee required additional information, to be provided by 2007, on the pharmacological effects of steviol glycosides in humans. These studies should involve repeated exposure to dietary and therapeutic doses, in normotensive and hypotensive individuals and in insulin-dependent and insulin-independent diabetics.