Stevia: Stevioside

Absorption, distribution and metabolism

In the rat, stevioside (125 mg/kg; p.o.) has a half-life of 24 hour, and is largely excreted in the feces in the form of steviol. Other metabolites include steviolbioside. In this species, at least, metabolism appears to be mediated primarily by the gut microflora. Thus, [17-14C] stevioside is converted to steviol by suspensions of rat intestinal microflora. Conversion is complete within two days.

The distribution of a derivative, [131I]iodostevioside (position of the label not reported), has been studied in rats following i.v. administration. Radioactivity rapidly accumulated first in the small intestine and then in the liver. Within two hours, 52% of the radioactivity administered appeared in the bile. The largest biliary component was [131I]iodosteviol (47% of total radioactivity), followed by [131I]iodostevioside (37%) and an unidentified metabolite (15%).

Non-enzymatic conversion of stevioside to steviol does not occur. Acid hydrolysis yields isosteviol, while incubation for up to three months under conditions ranging pH 2–8 and 5 to 90 °C does not result in detectable formation of steviol. Stevioside appears to be poorly transported across the cell membrane. No uptake was observed in suspensions of human red blood cells. The volume of distribution of stevioside was identical to that of [14C] sucrose. In the isolated, perfused rat liver, no metabolism of stevioside occurs over a two hour period at concentrations of 0.2 mM and 0.5 mM.

Lethality and whole animal toxicity

In whole animal studies, stevioside has shown low toxicity. From studies on crude Stevia extracts, it can be calculated that the acute LD50 of stevioside (i.p.) in rats is greater than 1700 mg/kg body weight. For oral administration in mice, an LD50 of 8.2 g/ kg has been reported in one study, and an LD50 of >15 g/kg has been calculated in another (for stevioside of 93–95% purity). In mice, no toxicity was detected following 2 g/kg of pure stevioside. No toxic effects were detected in mice two weeks following administration of doses up to 2 g/kg (by gastric intubation). More recently, stevioside has been fed to rats at up to 5% of the diet for 104 weeks with no evidence of carcinogenicity.

Yamada et al. () calculate a maximum no-effect level in rats of 550 mg/kg/day for a two-year toxicity and carcinogenicity study. Xili et al. () fed stevioside (85% pure) to rats at a dose of 800 mg/kg/day for two years. The maximum no-effect level of stevioside was equivalent to 1.2% of the diet, or 794 mg/kg/day. By extrapolating this figure to humans and allowing a safety factor of 100, these authors calculate an acceptable daily intake (ADI) of 7.94 mg/kg/ day. This is equivalent to 476 mg for a 60 kg adult or 160 mg for a 20 kg child. Assuming a mean daily intake of 50 g sucrose and a maximum replacement of 50% of this by stevioside, these authors calculate the ‘likely maximum intake’ of stevioside by humans to be 125 mg/day, or approximately 2 mg/kg (with the assumption that 1 mg stevioside replaces 200 mg sucrose). This is well below their calculated ADI, and so they conclude that stevioside is safe for human use. However, the assumptions in this paper seem dubious for a North American population. The ADI of 160 mg for a 20 kg child is equivalent to 32 g sucrose. A so-called ‘thirst quencher’ on the table as I write contains 14 g of sucrose/dextrose per 240 ml portion, while the various energy drinks in my collection contain up to 13 g sugar per 100 ml. With 50% replacement of sucrose by stevioside, many children in America, therefore, could potentially ingest well in excess of 2 mg/kg stevioside.

Pharmacological actions

Energy metabolism

Stevioside interferes with oxidative phosphorylation in isolated mitochondria, with 50% inhibition seen at 1.2 mM. This inhibition appears to be due to disruption of adenine dinucleotide translocation, a necessary process in shuttling high energy phosphate groups generated in mitochondria to their sites of consumption in the cell. In rat liver mitochondria, stevioside (5 mM) abolished coupled respiration (i.e. the increased O2 uptake in the presence of ADP). It also inhibited the stimulation of mitochondrial ATPase induced by the uncoupling agent, 2, 4-dinitrophenol. An inhibition of 50% was achieved at 1.2mM stevioside. At 1.5 mM, it is without effect on glutamate dehydrogenase activity of rat or bovine liver mitochondria. Stevioside is much less potent, however, than atractyloside, a compound of related structure, but having a free carboxyl group. At 1 µM, this latter compound produces 50% inhibition of oxidative phosphorylation.

Some of the mitochondrial actions of stevioside are summarized in Table “Inhibitory constants (half-maximal effects: I50) for Stevia rebaudiana glycosides and aglycones on mitochondrial activities”. High levels of stevioside (1.5 g/kg body weight; s.c.) caused severe disruption of mitochondrial cristae in kidney tubules. The mitochondrial actions of stevioside are seen only on isolated organelles, and are not observed in intact cells. This suggests that stevioside does not permeate cell membranes, although this point does not appear to have been directly examined apart from the distribution studies discussed above.

Effects on energy metabolism that do not involve mitochondria have been examined in cells lacking these organelles. Erythrocytes rely on glycolysis for ATP production. Stevioside has little effect on such cells.

Carbohydrate metabolism

The effect of stevioside on energy metabolism in isolated mitochondria has spurred the search for related effects on intact cells. At a concentration of 3 mM, however, stevioside was without effect on gluconeogenesis or O2 uptake in renal cortical tubules.

At the whole organ level, stevioside inhibits the monosaccharide transporter in the isolated, perfused rat liver. The transporter carries glucose, fructose and galactose in both directions. Stevioside (0.8 mM) halves the transport rate of glucose into the liver. At 1.5 mM stevioside, there is 73% inhibition of 1 mM glucose uptake, combined with inhibition of fructose metabolism. Stevioside also inhibits hepatic release of glucose. In livers undergoing glycogenolysis, the presence of stevioside leads to an increase in the intracellular:extracellular concentration gradient of glucose. Both the Km and Vmax of monosaccharide transport are altered, suggesting a mixed type of inhibition. In the rat kidney, also, stevioside given by in vivo infusion (8 mg/kg/h, or higher) decreased the renal tubular resorption of glucose.

In hamsters fed stevioside (2.5 g/kg/day) for 12 weeks, glucose absorption was inhibited. Doses of 1 g/kg/day were without effect. The effect was attributed to two actions: (i) a decrease in intestinal Na+/K+-ATPase activity; and (ii) a decreased absorptive area in the intestines. Body weight of the hamsters decreased over the course of the study. In contrast, stevioside (5 mM) has been reported to have no inhibitory effect on glucose absorption from the rat jejunum in vitro ().

No change in blood sugar levels in rats was observed following administration of stevioside (7% of diet) for 56 days. A similar result was obtained from a study in which stevioside (75–150 mg/kg/day) was fed for 30 days. However, an increase in plasma glucose in rats was found when stevioside was given by intravenous infusion. A loading dose of 100 mg/kg body weight followed by an infusion of 100 mg/kg body weight/hour for one hour resulted in a significant increase in plasma glucose level. The increase reached 47% at 200 mg/kg body weight loading and infusion. There was no change in plasma insulin levels. The authors attribute their findings to an increased glucose transport into cells.

In 24-hour-fasted rats, stevioside (0.2 mM) given orally with fructose (0.2 mM) as a gluconeogenic substrate led to increased glycogen deposition in the liver. In fasted rats, stevioside in the drinking water (1 or 2 mM) also increased glycogen deposition in the liver in the absence of a gluconeogenic substrate. It is difficult to reconcile these findings with inhibition of monosaccharide transport. Conversely, in the same species, 0.1% stevioside in a high carbohydrate diet decreased liver glycogen levels, but was without effect in a high fat diet. Clarification is needed for the effects of stevioside on carbohydrate metabolism in the intact animal.

Effects on blood pressure and renal function

Various effects have been claimed for stevioside on kidney function and blood pressure regulation. It lowers mean arterial blood pressure, decreases renal vascular resistance, produces diuresis, and increases fractional excretion of Na+ and K+. The lack of effect on glomerular filtration rate implies that stevioside vasodilates both afferent and efferent arterioles.

One group has studied renal effects by using an in vivo priming dose of 8 or 16 mg/kg, followed by an infusion at the rate of 8 or 16 mg/kg/h. The higher dose led to a marked fall in blood pressure in anesthetized rats (from 110 to 72 mm Hg). All the above effects observed in rats were blocked by indomethacin, an inhibitor of cyclooxygenase, an enzyme involved in prostaglandin synthesis. This suggests that the actions of stevioside are mediated via a prostaglandin-dependent mechanism.The Ca2+-channel antagonist, verapamil, increased the vasodilating effects of stevioside, while calcium chloride antagonized them. This has been interpreted as indicating that the effects of stevioside are due to Ca+channel blockade. However, one suspects the same effects of verapamil and calcium chloride would be seen regardless of the mechanism by which stevioside is acting.

In the Goldblatt hypertensive rat, stevioside given by the same protocol (16 mg/kg priming dose; 16 mg/kg/h infusion) also decreased mean arterial pressure and increased renal plasma flow. Glomerular filtration rate is also increased, implying a preferential vasodilation of afferent arterioles.

When administered in vivo, stevioside (1.5 g/kg, s.c.) in rats decreased the ability of isolated renal cortical tissue to accumulate p-aminohippuric acid nine hours later. Infusions (i.v.) of much lower levels, down to 8 mg/kg/h, also increased clearance of p-aminohippuric acid and glucose.

Stevioside itself had a clearance less than that of p-aminohippuric acid but greater than that of inulin. Inulin clearance is a measure of glomerular filtration rate. This finding, therefore, suggests that the glycoside is secreted by the renal tubular epithelium.

Nephrotoxic effects of stevioside have been observed in the rat following a single injection of 1.5 g/kg, s.c..This is a dose equivalent to about 250 times the average daily intake of human consumers. Blood urea nitrogen began to rise three-hour later, an indication of inability to excrete nitrogen. Blood levels were highest at nine hour (52.4 ± 2.9 mg/% compared with 18.7 ± 0.7 in control animals), and were still elevated 48 hours post-injection. Other biochemical disturbances nine hours post-injection were elevated plasma creatinine, increased urinary glucose (from 4.9 to 47.6 mg/%), and increases in urinary levels of two enzymes: γ-glutamyltranspeptidase (from 0.12 to 0.99 IU/ml) and alkaline phosphatase. In association with these changes were electron microscopic indications of degeneration of kidney tubules with severe disruption of mitochondrial cristae. It is possible that these effects are a consequence of hydrolysis to steviol. Similar nephrotoxic effects have also been reported by others, both in the rat and hamster, although I have not had access to the full report.

These findings indicate that at the doses and in the species used, stevioside has vasoactive properties.

Reproductive and teratogenic effects

Stevioside (0.15–3% of diet) was fed to male rats for 60 days before mating and to female rats for 14 days prior to mating and for the first seven days of gestation. No effects were seen on either mating performance or fertility. The group receiving the highest exposure had a slight retardation in weight gain. Given by gavage once a day through days 6 through 16 of pregnancy, stevioside (0–10 00 mg/kg/day) produced no fetal malformations or other toxic effects.

Chromosomal and mutagenic effects

Chromosomal abnormalities are only seen with stevioside at high concentrations. In Chinese hamster D-6 cells cultured with various concentrations of stevioside for 28 hours, aberrations were only seen at concentrations of 2% (24.8 mM) or higher. The commonest alteration was an increase in the number of gaps and interchanges. Sister chromatid exchanges also increased in a dose-dependent manner. In a Chinese hamster fibroblast cell line, stevioside (85% pure; 14.9 mM) did not induce chromosomal aberrations.

No chromosomal effects of stevioside were noted in cultured human lymphocytes. No mutagenic effects were noted in six in vitro (four bacterial and two mammalian) and one in vivo screens.

When tested against two strains of Salmonella typhimurium, no mutagenic effects of stevioside were noted at a concentration of 25 mg/plate. At 50 mg/plate, mutagenicity was shown against one strain. No mutagenic action against S. typhimurium was noted in the presence of an activating system isolated from Aroclor-pretreated rats.


Stevioside has a number of pharmacological actions, including disruption of oxidative phosphorylation, inhibition of monosaccharide transport and impairment of kidney function. Some of these actions have been demonstrated on intact animals or isolated organs, others only on subcellular preparations. A number of authors have suggested therapeutic applications of certain of these actions, particularly of the hypotensive effect. In terms of use as a preferably inert food additive, possession of a potentially therapeutic property implies the potential for a toxic effect. For example, Stevia preparations have been proposed as being of benefit in weight control. In other circumstances, a decreased weight gain accompanying the use of a food additive might be considered a toxic response.

In an example of having their cake and wanting to eat it, the authors of one of the major whole animal studies of the safety of stevioside conclude that the material is safe and has no adverse effects. However, they go on to conclude that stevioside may have ‘therapeutic value in the treatment of patients with diabetes-related obesity , hypertension or cardiac disease’ (what is meant by the latter being undefined). If an agent can lower blood pressure or affect other aspects of cardiovascular functioning, it can clearly produce adverse or unwanted effects. Pharmacology and toxicology are the two faces of the same coin.

The absorption, distribution, metabolism and excretion of stevioside have been poorly studied, particularly in non-rodent species.

In view of the toxicity of the aglycone, steviol, more information is needed on the ability of mammals, or their intestinal microflora, to convert stevioside to steviol. Are human microflora capable of this transformation? Do certain disease states or antibiotic regimens alter microflora and their ability to perform this transformation? In view of the toxicity associated with related compounds, such as atractyloside, having a free carboxyl group at the C-4 position, the mammalian ability to convert stevioside to steviolbioside also needs to be investigated. This glycoside has been postulated to be an intermediate in the mammalian conversion of stevioside to steviol.

In in vitro preparations, stevioside can interfere with oxidative phosphorylation and energy metabolism in mitochondria. However, the ability of this glycoside to penetrate cell membranes does not appear to have been sufficiently studied.