Stevia (from the plant stevia rebaudiana in the family Compositae) is a small perrential herb with a history and current usage of being a sweetening agent. There are 154 plants within the stevia family with all plants being sweet but rebaudiana having the highest sweetness levels. The usage of stevia as a sweetener relative to table sugar (sucrose) is said to range from 200-300 sweeter (on a gram per gram basis)] or simply be 250-fold more potent.
The main bioactive in stevia is the diterpene known as steviol, although it itself is in low concentrations with the majority of the plant containing glycosides of steviol (steviol bound to sugars); rebaudioside A is also used as a sweetening agent and may be causative of the reported bitter aftertaste via two particular bitter receptors (hTAS2R4 and hTAS2R14). On a molecular level, stevioside is approximately 300-fold more potent than sucrose whereas rebaudioside A is 400-fold sweeter.
Usage of steviol glycosides for sweetening has a GRAS (Generally recognized as safe) rating in the US as appears to have an adequate daily intake (ADI) of 25mg/kg (following 100-fold safety factor, commonly seen in ADI values) in rats which is around 7.9mg/kg in humans.
Stevia is a leaf that it, and its extracts, are used as sweetening agents as on a gram per gram basis they are about 200-300x as sweet as sugar. Stevia may have a slightly bitter aftertaste, and with infrequent usage in food products is generally recognized as safe
Steviol (low conentrations, 5.9+/-0.8μg/g or 0.00059%) and its glycosides stevioside (4-13% dry weight) and steviolbioside (thought to be less than 1%, although the detection of steviolbioside may be an artefact of extraction techniques)
Rebaudioside A (2-4%, although 1.62-7.27% has been reported), C (1-3%) and the glycosides of rebaudiosides B, D, E, and F as well as rubusoside (all less than 1%). Similar to steviolbioside, rebaudioside B may be an artefact
Chlorophyll a (0.2mg/g) and b (0.3mg/g)
There appears to be a protein content up to 15.523+/-1.877% which includes (in reference to total dry weight):
Proline at 1.729+/-0.581%
Histidine at 0.343+/-0.045%
Lysine at 1.062+/-0.151%
Phenylalanine at 0.887+/-0.116%
Isoleucine at 0.723+/-0.095%
Valine at 0.948+/-0.108%
Alanine at 0.954+/-0.115%
Glycine at 0.856+/-0.100%
Serine at 1.027+/-0.339%
Threonine at 0.754+/-0.099%
Aspartic acid at 1.727+/-0.278%
Similar to all plants, there can be a large degree of variability due to growing conditions and cultivar with a range of total steviosides having been noted to be 4-20% and extraction techniques further influencing the steviol glycoside content. The total polyphenolic content of stevia leaves is around 91mg/g (9.1% dry weight) with the flavonoid component around 23mg/g (2.3%) of the total leaf extract by dry weight which can be elevated to 5.7% (56.73mg/g) in the aqueous extract.
Stevia is mostly a collection of glycosides of the diterpene known as steviol, which is in remarkably low amounts by itself as it appears to be heavily sequestered by sugars (wherein the glycosides stevioside and rebaudioside A may convert to steviol after bacterial fermentation). There are some other diterpenes and flavonoids, although it is not known if these molecules exert much of the bioactivity seen with stevia supplementation
Stevioside and Rebaudioside A appear to be resistant to degradation by sunlight when in solution, and when forcing degradation (80C temperature for 72 hours in a carbonated drink) although both stevioside and rebaudioside A experienced degradation stevioside appears to be less stable. If not forced, these glycosides appear quite stable with a report existing of a 62 year old leaf (conditions of storage not reported) still having a sweet taste.
Rebaudioside A in isolation (sold under the brand name of Rebiana) is an odourless white to off-white powder that appears to be water soluble and highly stable in powder form (up to 2 years assuming ambient humidity and temperature) and stable in solution with pH values between 4-8, with stability dropping when pH drops below 2. When stored at 40-60C for 1-14 days, rebaudioside A appears stable when in the matrix of a non-carbonated sweetened beverage (pH 7.6) and chocolate milk (pH 10.1) whereas it experienced a small degree of degradation in a carbonated beverage (pH 3.2) with another study using more drastic experimental conditions (carbonated beverage stored at 80C for 72 hours) noting a 54% loss of rebaudioside A.
The steviol glycosides appear fairly stable when in powder form and in solution, and although they appear photoresistant they may be degraded with excessive heat or an acidic pH
Following oral ingestion of stevia, the molecule known as steviol glucuronide (steviol glycosides metabolized into free steviol, and then glucuronidated by the body) appears in serum with a prolonged Tmax of 12 and 8 hours (rebaudioside A and stevioside, respectively), Cmax values of 1472ng/mL and 1886ng/mL, and an AUC of 30,788ng/h/L and 34,090ng/h/L.
Both rebaudioside A and stevioside can increase serum concentrations of steviol glucuronide
Steviol appears to be glucuronidated in the body into a metabolite known as steviol glucuronide.
Glycosides are sometimes not well absorbed in humans, and instead subject to intestinal fermentation by colonic bacteria. Using bacteria taken from human feces, stevioside and rebaudioside A are both noted to be hydrolyzed into free steviol (similar to previous studies conducted on rat intestinal bacteria) and steviol does not appear to be further metabolized. When assessing the composition of fecal bacterial cultures, there does not appear to be any significant influence on stevioside or rebaudioside A.
Steviol glycosides can be metabolized into free steviol in the intestines by colonic microflora, and steviol does not appear to be metabolized any further (bacterial population does not seem altered in a beneficial nor detrimental manner)
In lean and obese persons given either a test meal containing sucralose (490kcal) or one of two meals sweetened with low calorie replacements (290kcal of food alongside either aspartame or stevia) noted that both test meals were equally satiating as the higher caloric meal and no overall compensation was noted when measuring calories eaten; stevia and aspartame were not significantly different.
In serum starved PC12 cells (which undergo apoptosis via DNA fragmentation), the incubation of these neurons with up to 10mcg/mL stevioside (a dose which is not inhernetly cytotoxic) appeared to augment apoptosis induced by serum starvation at the concentrations of 100-1000ng/mL.
Isosteviol appears to have a neuroprotective effect in a rat model of cerebral occlusion, where 5-20mg/kg intravenous isosteviol exerted protective effects secondary to its antiinflammatory properties (NF-kB inhibition) in a dose dependent manner and with 20mg/kg isosteviol being as effective as 5mg/kg nimodipine.
Some neuroprotective effects associated with stevia consumption, difficult to assess practical relevance of the above information
4 weeks consumption of 1000mg of isolated rebaudioside A failed to significantly modify systolic or diastolic blood pressure and 750mg stevioside in another trial lasting for 3 months similarly failed to modify blood pressure overall.
One study assessing tolerability of 1500mg stevioside (thrice daily dosing of 500mg) as an end point noted that 2 years of routine usage of stevioside was associated with a decrease in both systolic and diastolic blood pressure by 6.7% and 6.4%, respectively; this study also noted that treatment was associated with a reduced rate of left ventricular hypertrophy with 11.5% of stevia and 34.0% of placebo. This is similar to a year long trial in hypertensives given stevioside at 250mg thrice daily (750mg total) who experienced a decrease in blood pressure to a slightly higher magnitude (8.1% and 13.8% systolic and diastolic) relative to placebo at the third month which maintained its magnitude over the subsequent nine months.
No apparent significant effects of steviosides at standard doses on blood pressure in persons with normal blood pressure, two studies to support a decrease in blood pressure in hypertensives with prolonged usage
Compared to regular cookies (240 kcal; 50 g CHO), consuming stevia cookies (240 kcal; 50 g CHO; 3 g ground stevia leaves) has no significant effect on the 2-hour postprandial blood glucose response (insulin response not reported).
When rats are fed stevia leaf (4% of feed) for 4 weeks prior to streptozotocin injections (to induce diabetes) and then sacrifed a week later noted that stevia was able to attenuate the increase in oxidation seen with steptozotocin control rats (this also applied to the polyphenolics in isolation, but not the fiber content) and attenuated the rise in blood glucose seen in diabetic controls, with polyphenolics form the stevia leaf outperforming the leaf extract.
Isolated rebaudioside A at 50-200mg/kg in already diabetic rats for 45 days was able to somewhat preserve the loss of serum insulin and glucose in diabetic rats (200mg/kg being slightly less effective than 600mcg/kg glibenclamide) with 200mg/kg failing to have any significant influence on non-diabetic rats. An attenuation in liver enzymes that are induced during diabetes (G6P, F16BP, Hexokinase) was also noted with 200mg/kg being comparable to 600mcg/kg glibenclamide and both lower (200-400mg/kg) and higher (2,000mg/kg) doses of stevia have shown an ability to reduce blood glucose in alloxan-induced diabetic rats which are thought to be secondary to aiding pancreatic function.
Stevia components appear to reduce blood glucose in diabetic rat models, although at a dose somewhat above the ADI (not in the dosage range where toxicity is noted) and by general mechanisms. Relative to other supplemental compounds, it doesn't appear to be overly remarkable nor unique
One study has been conducted in diabetic humans, where a single acute dose of stevioside (1,000mg) was able to reduce the postprandial AUC of glucose by 18% relative to control (1g corn starch) and appeared to benefit the insulin:glucose ratio in serum by 40%.
A single acute study on stevioside, where glucose in serum following a meal was reduced
Secondary to its anti-inflammatory actions (inhibition of NF-kB, which is known to cause muscle hypertrophy promoting effects in animals), 10mg/kg of stevioside daily to rats (orally) for one week prior to induce toxicity to skeletal muscle (via cardiotoxin) noted that stevioside was associated with increase muscle regeneration rates via increasing satellite cell recruitment and increased functional capacity of injured muscle 7 days post injury.
Some muscle regenerative effects secondary to NF-kB inhibition noted, which may not be a mechanism unique to stevia (NF-kB inhibitors are common amongst plants)
In mice, 12.5-50mg/kg stevioside oral ingestion and hour prior to LPS injections was able to attenuate the increase in lung immune cell (neutrophil and macrophage) count at 25-50mg/kg at the same potency as 5mg/kg dexamethasone (trended to be weaker) and similarly reduced the levels of proinflammatory cytokines TNF-α, IL-1β, and IL-6 and nitric oxide/iNOS activity (similar to dexamethasone) as well as reducing the induction seen in COX2; these effects were thought to be secondary to reducing NF-kB activation in response to LPS. This has been noted previously in macrophages, THP-1 immune cells, as well as intestinal cells where stevioside inhibited NF-kB activity and is thought to be secondary to suppressing LPS-induced MAPK activation (which then proceeds to influene NF-kB and inflammation via the MAPK/NF-kB pathway).
Stevioside appears to possess anti-inflammatory properties in rats following oral ingestion of somewhat feasible doses (human equivalent of a rat 25-50mg/kg is 4-8mg/kg and, for a 150lb person, 272-545mg) which is currently thought to be due to inhibiting MAPK
When testing stevia leaf extracts against bacterial strains involved in dental caries (12 strains of Streptococcus and 4 strains of Lactobacillus) noted a minimum inhibitory concentration between 30-120mg/mL on all bacteria with a hexane extraction being most effective.
May have some anti-cavity effects via its anti-bacterial properties, but this has not yet been tested in vivo and occurs at a fairly high oral dose not common to sweeteners
A molecule in stevia known as dihydroisosteviol has once shown antiandrogenic action by interfering with the effects of testosterone injections in a chick comb (at 3mg) without significantly affecting the ability of testosterone to influence testicular function. Stevioside may also displace DHT from the androgen receptor (conference presentation cited via here) at concentrations ot 7-15µM, which has been argued that it may not be practically relevant as even oral doses of very high (6.7g/kg in rats) for 60 days failed to alter androgen binding in prostatic tissue.
When rats are fed stevia (about 6.7g/kg) daily for 60 days, testosterone can be reduced by approximately 40% (data derived from graph) with no influence on luteinizing hormone.
Very high doses of stevia may reduce testosterone levels (secondary to testicular damage) although the direct binding of stevia compounds to the androgen receptor does not seem to be practically relevant due to even higher concentrations required
Dihydroisosteviol is known to activate AMPK in intestinal cells which underlies its ability to inhibit CFTR transport mediated chloride secretion into the intestines. AMPK colocalizes with CFTR and its activation in general, in intestinal cells, suppresses its actions. This same study found a failure to activate AMPK in thyroid cells, suggesting a site-specific effect.
One month of ingesting stevia leaf (4% of feed) to rats for 5 weeks noted that the increase in urinary protein and creatinine seen during diabetes (chemically induced at week 4) was attenuated, but not abolished. Higher doses (2.67g in rats) seem to increase renal blood flow and glomerular filtration rate and decrease the resorption of glucose from renal tubules when injected at the rate of 1-3mg/kg/h, but unable to do so at 0.5mg/kg/h.
A diuretic effect has been noted with doses of stevia higher than used in sweetening when tested in rats with an intravenous infusion of 0.5mg/kg/h in rats being ineffective on this urinary parameter.
Some kidney protective and diuretic effects associated with doses of stevia higher than that used in sweetening, practical relevance of the above is uncertain in humans (and although there is protection noted, it is not to a remarkable degree)
In rats fed 4% of the diet as stevia leaves for 5 weeks, the leafs themselves as well as isolated polyphenolics were able to inhibit the increase of liver enzymes seen with diabetes.
In prepubertal male rats fed stevia (about 6.7g/kg) for 60 days there are reductions noted in the weight of the Cauda epididymidis (52%), testes (50%), and seminal vesicle (51%) with no influence on overall body weight or prostatic weight. This same dose has elsewhere been noted to reduce the weight of the seminal vesicle by 60% in rats although the testes were unaffected in this study.
Spermatozoa has also been reduced at 6.7g/kg for 60 days to 68% of control.
High doses of stevia can impair testicular and seminal function in male rats
Stevioside has been noted to induce breast cancer cell (MCF-7) apoptosis with an IC50 between 10-20μM after 72 hours, this apoptosis being mediated via a mitochondrial-dependent ROS generation pathway. 10μM of stevioside over 72 hours is associated with 71.1% cell population undergoing apoptosis and an increase in sub G0/G1 cell phase, and concentrations as low as 2.5μM were somewhat effective while concentrations over 20μM toxic.
This study also noted that Nrf2 expression was increased at the end of 72 hours of incubation with a downregulation in Keap1 (Nrf2 antagonist) activity.
Stevioside, Steviol, and Isostevioside appear to have anti-cancer effects in a mouse model of skin cancer (DMBA or peroxynitrate induction followed by TPA augmentation) by reducing papilloma count after 9 weeks (topical application of 85nmol of the molecule in question at the site of skin tumor induction) by 44.4-46.2%, a potency comparable to curcumin (45.1%).
The adequate daily intake (ADI) of stevia appears to be between 7.9mg/kg and 25mg/kg with the latter deduced from a No Observable Adverse Effect Limited (NOAEL; the highest dose without adverse effects) of over 2,500mg/kg in rats (over 3 months) and hamsters over a few generations while the former is an estimated human ADI based on rat data.
The acute LD50 value (amount required to induce death in half of the studied population) of 90% pure stevia appears to be in the range of 5,200-6,100mg/kg in hamsters and above 15,000mg/kg in rats and mice. As mentioned earlier in sections on fertility, doses at around 6,700mg/kg can suppress testicular function in otherwise healthy rats despite being under the LD50 while reductions in female fertility have been noted to occur at lower doses following ingestion of 10mL water with 5% stevia (500mg overall; extrapolated from data where a 5% decoction is made from 15g of the leaves and 300mL water in traditional medicine) which correlates to a human dose of approximately 320mg/kg (assuming rats weight 250g). This infertility in female rats appears to persist for 50-60 days after cessation of stevia an appears to have been noted in one other instance (Portella Nunes, B.A., Pereira, N.A 1998; found via here but not located online) but has elsewhere failed to have an effect with up to 10% concentration in 1mL of water.
For isolated rebaudioside A, there does not appear to be any adverse effects with 90 days supplementation of up to 2000mg/kg in rats.
Unlike other sweeteners, stevia has both biologically relevant effects as well as possible toxicity with overdoses (as evidenced by a detectable LD50 in mammals). There appear to be anti-fertility effects, and although these effects in males are not a concern due to occurring at high doses it appears female infertility occurs at lower concentrations.
Stevioside (50mg per plate in the Ames test) has either been found to not exert genotoxic effects in Salmonella typhimurium TA98 and TA100 or weak genotoxic effects in TA98, one study using a battery of genotoxicity tests failed to determine a significant influence of stevioside.
Steviol (2mg per plate) has also been detected to have no genotoxic effects in one study and has failed to exert any abnormal effects in lymphocytes isolated from healthy volunteers at a concentration of 0.1-0.2mg/mL. In a battery of tests, steviol has been noted to exert dose-dependent genotoxic effects in a forward mutation assay (TM677) without having any apparent effect in reverse mutation assays up to 5mg/mL and a subsequent test in TM677 confirmed these effects. The potency of steviol, relative to the known mutagen Furylfuramide (AF-2) is about 0.004% as potent (and to achieve the same genotoxic effects, the authors calculated that approximately 3000 cups of coffee would need to be sweetened).
One study in male wistar rats orally with 4mg/mL (88.62% purity steviosides) of solution in the drinking water for 45 days noted increased genotoxicity in peripheral blood cells as assessed by comet assay with effects also detected in the liver, spleen, and brain; no lethality was noted in this study The study failed to disclose the overall amount of stevioside consumed, but elsewhere has been estimated at 80mg daily (assuming 20mL water consumption) which equates to 500mg/kg in these rats and has been criticized for its methodology such as not having a control or dose-comparative groups or controlling feed quality. Other studies in mammals failed to note an effect of high doses (up to 2000mg/kg) of stevia acutely (24 hours) in mice and according to the Joint Expert Committee on Food Additives (JECFA) stevia failed to be carcinogenic up to 2000mg/kg and 2400mg/kg in male and female rats, respectively (cited vicariously through, as this was reported via conference).
As for rebaudioside A, there does not appear to be any significant influences on genotoxicity or carcinogenesis following ingestion of up to 100,000ppm and a quarter this dose for up to two generations or 750-2000mg/kg bodyweight in mice.
Although steviosides and steviol are not absolutely genotoxic, there is some evidence that in the genotoxic test of the bacteria Salmonella typhimurium TM677 forward mutation that steviol and its metabolites are genotoxic. Relative to known mutagens, steviol appears to be pretty weak and likely not practically relevant. Animal evidence at this moment in time is limited and contradictory
In the context of oxidative damage, stevia extracts (0.1mg/mL) can exhibit genoprotective effects in vitro secondary to its antioxidant properties.