Stevia Leaf to Stevia Sweetener: Exploring Its Science, Benefits, and Future Potential

Abstract

Steviol glycoside sweeteners are extracted and purified from the Stevia rebaudiana Bertoni plant, a member of the Asteraceae (Compositae) family that is native to South America, where it has been used for its sweet properties for hundreds of years. With continued increasing rates of obesity, diabetes, and other related comorbidities, in conjunction with global public policies calling for reductions in sugar intake as a means to help curb these issues, low- and no-calorie sweeteners (LNCSs, also known as high-potency sweeteners) such as stevia are gaining interest among consumers and food manufacturers. This appeal is related to stevia being plant-based, zero calorie and with a sweet taste that is 50–350 times sweeter than sugar, making it an excellent choice for use in sugar- and calorie-reduced food and beverage products. Despite the fact that the safety of stevia has been affirmed by several food regulatory and safety authorities around the world, insufficient education about stevia's safety and benefits, including continuing concern with regard to the safety of LNCSs in general, deters health professionals and consumers from recommending or using stevia. Therefore, the aim of this review and the stevia symposium that preceded this review at the ASN's annual conference in 2017 was to examine, in a comprehensive manner, the state of the science for stevia, its safety and potential health benefits, and future research and application. Topics covered included metabolism, safety and acceptable intake, dietary exposure, impact on blood glucose and insulin concentrations, energy intake and weight management, blood pressure, dental caries, naturality and processing, taste and sensory properties, regulatory status, consumer insights, and market trends. Data for stevia are limited in the case of energy intake and weight management as well as for the gut microbiome; therefore, the broader literature on LNCSs was reviewed at the symposium and therefore is also included in this review.

Introduction

Stevia rebaudiana Bertoni is a small perennial shrub of the Asteraceae (Compositae) family that is native to Paraguay, Brazil, and Argentina. The leaves of this plant have been used by indigenous people for centuries in medicines and to sweeten drinks such as maté, a green herbal tea (1–3). The plant was first brought to the attention of the rest of the world by the botanist Moises Santiago Bertoni in 1887, who learned of its properties from the Paraguayan Indians (1, 3). The chemical characterization of the natural constituents of the plant known as steviol glycosides, which are responsible for its distinct sweet taste, was not identified until 1931 when 2 French chemists, Bridel and Lavielle, isolated stevioside, a primary steviol glycoside from stevia leaves (1). Japan was the first country to commercialize and use crude, unpurified extracts of S. rebaudiana in the 1970s on a large scale (2). Its use eventually spread to several countries in Asia and Latin America (4). In the 1990s, stevia extract was available in the United States as a dietary supplement in health food stores; however, early formulations were known to have a licorice flavor with a sweet or bitter aftertaste, which limited their widespread commercial development (2, 5). The presence of essential oils, tannins, and flavonoids in the crude extracts was partly responsible for some of the off tastes; hence, efforts were made to purify extracts and chemically characterize steviol glycosides (5).

After the isolation of stevioside, several other steviol glycosides, such as rebaudiosides (Reb) A, B, C, D, and E and dulcoside A, were identified and isolated from stevia leaves (6). Generally, the most abundant steviol glycosides in stevia leaves are stevioside (4–13% wt:wt), Reb A (2–4%), and Reb C (1–2% wt:wt) (7, 8). To date, >40 steviol glycosides have been identified (e.g., Reb F, G, H, I, J, K, L, M, N, O, and Q; stevioside A, D, and E, etc.) (9–12). Most of the steviol glycosides derived from the plant are 4-ring diterpenes that have a backbone of 13-hydroxy-ent-kaur-16-en-19-oic acid, known as steviol (1, 12). The various glycosides differ only in the number and type of monosaccharides attached at the R1 (OH) and R2 (H) position of the aglycone steviol. Glucose, fructose, rhamnose, xylose, and deoxyglucose are examples of sugars that are attached to the steviol backbone (12). The 2 primary steviol glycosides, stevioside and Reb A, differ only by 1 glucose moiety at R1; stevioside has 2 glucose molecules, whereas Reb A has 3.

The stevia plant is now commercially cultivated in Argentina, Brazil, Columbia, Paraguay, China, Japan, Malaysia, South Korea, Vietnam, Israel, Australia, Kenya, and the United States. High-purity steviol glycosides are approved as sweeteners by all major regulatory authorities across the globe, and >150 countries have approved or adopted its use in foods and beverages. Reb A was the first commercial steviol glycoside launched in the marketplace (13).

Metabolism of steviol glycosides

The absorption, metabolism, and excretion of steviol glycosides have been extensively reviewed by multiple scientific authorities and experts, including the European Food Safety Authority (EFSA) (14), and recently by Magnuson et al. (15). Steviol glycosides are undigested in the upper gastrointestinal tract. They are hydrolyzed or degraded only when they come into contact with microbiota in the colon that cleave the glycosidic linkages, removing the sugar moieties, leaving behind the steviol backbone that is absorbed systemically, glucuronidated in the liver, and excreted via urine in humans and via feces in rats (15).

In vitro studies show that human saliva, salivary α-amylase, pepsin, pancreatin, and pancreatic α-amylase, as well as jejunal brush border enzymes of mice, rats, and hamsters, are not able to hydrolyze the glycosidic bonds present in stevioside (16). However, the gut microbiota of humans, rodents, and hamsters are able to degrade stevioside to steviol (16). Incubation of stevioside and Reb A with human fecal microbiota showed that both were completely hydrolyzed to steviol in 10 and 24 h, respectively (4, 17). The released sugar moieties are not absorbed and are most likely quickly utilized by the gut microbes as an energy source, thus making it a zero-calorie sweetener (2). An in vitro model of the intestinal barrier has shown that the transport of stevioside and Reb A through the monolayers is very low, whereas the absorptive transport of steviol is high, suggesting that steviol is not metabolized by gut microbiota and is absorbed from the intestine (18). Bacteroides species are primarily responsible for the hydrolysis of steviol glycosides in the gut via their β-glucosidase activity (17).

Evidence from in vitro investigations are consistent with human metabolism studies that showed no detectable presence of the glycosides in plasma, suggesting no uptake from the gut and little or no stevioside or Reb A in urine or feces (19–22). These studies also showed that steviol is absorbed quickly and transported to the liver where it is conjugated with glucuronic acid to form steviol glucuronide, which, in humans, is excreted in urine (19–22). Figure 1 summarizes the absorption, metabolism, and excretion pathway of steviol glycosides in humans.

Wheeler et al. (21) compared the pharmacokinetics and metabolism of stevioside and Reb A in healthy adults over a 72-h period. Peak plasma concentrations occurred at 8 and 12 h for stevioside and Reb A, respectively, and a half-life (t_1/2) of 14–16 h was observed for both. Intake of Reb A resulted in significantly lower steviol glucuronide concentrations (59%) than after stevioside (62%) consumption. The differences in steviol glucuronide concentrations are attributed to the simpler structure and faster bacterial degradation of stevioside compared with Reb A. Fecal recovery of steviol accounted for ∼5% of the original dose for both compounds. The pharmacokinetic analyses showed that stevioside and Reb A undergo similar metabolic and elimination processes in humans.

Most of the earlier studies on steviol glycoside metabolism were on Reb A or stevioside (a.k.a. primary or major glycosides). However, the similarities in the microbial metabolism of several steviol glycosides were confirmed in in vitro studies of pooled human fecal homogenates of healthy male and female Asian and white subjects (12, 23). Reb A, B, C, D, E, F, and M; dulcoside A (a.k.a. minor glycosides); and steviolbioside (an intermediate metabolite), which contain different sugar moieties (glucose, rhamnose, xylose, fructose, and deoxyglucose) and different linkage types [αβ (1–2), β-1, β (1–2), β (1–3), and β (1–6)], were all degraded to steviol within 24 to 48 h. No differences between male and female subjects or between ethnicities were observed. These data suggest that the different steviol glycosides have similar hydrolysis rates to that of Reb A and therefore would be expected to have steviol absorption rates, metabolism, and pharmacokinetics similar to Reb A. This was also confirmed in an animal model comparing the metabolism of Reb A and Reb D (24). These data show that both major and minor steviol glycosides appear to share a common metabolic fate.

Safety and acceptable daily intake of steviol glycosides

The safety of steviol glycosides from numerous toxicological, biological, and clinical studies has been reviewed in several publications (2, 7, 14, 25, 26). As described in the “Regulatory status” section of this review, all major global scientific and regulatory bodies have determined high-purity steviol glycosides to be safe for consumption by the general population. The majority of the regulatory approvals pertain to high-purity (≥95%) steviol glycosides. Unpurified crude extracts of stevia have been reported to cause adverse effects on fertility in animals (27, 28), which have not been observed with well-characterized, high-purity steviol glycosides approved for food and beverage use. Therefore, studies conducted with crude extracts have been determined to be not relevant to the safety assessment of high-purity steviol glycosides by knowledgeable scientific experts and regulatory authorities.

Potential effects of high-purity steviol glycosides on acute and long-term toxicity, reproductive and developmental toxicity, and carcinogenicity have been studied primarily in rodents but also in other animal models (29–34). Steviol glycosides are excreted primarily as steviol glucuronide in the urine in humans, whereas in rats, free steviol and steviol glucuronide are excreted primarily in the feces via the bile, with <3% appearing in the urine (2, 35). This interspecies difference is due to the lower molecular weight threshold for biliary excretion in rats than in humans (2). Although the elimination routes of steviol glycosides differ between humans and rats, this is of no toxicological significance because the metabolism and pharmacokinetics are similar in the 2 species (2). In other words, the majority of the tissues and cells of the body are exposed to similar concentrations of the same metabolites for a similar amount of time after consumption of steviol glycosides in both species, so the potential for development of a toxicological effect is similar, even though the final route of excretion is different. Therefore, the rat is an appropriate test animal to assess the safety of consumption of steviol glycosides, and toxicological data generated from rat studies are applicable to humans (2).

The Acceptable Daily Intake (ADI) is the amount of a substance that an individual can consume daily over a lifetime without any appreciable health risk. It is established by regulatory agencies on the basis of the results of toxicology testing. The No Observed Adverse Effect Level (NOAEL), which is the highest dose fed to animals in long-term studies with no adverse toxicological effect, is considered the basis of the ADI. The NOAEL is divided by safety factors (typically 100) to account for intra- and interspecies differences to ensure the ADI is safe for all potential consumers, including subgroups such as children. The current ADI for steviol glycosides is based on a toxicity and carcinogenicity study that tested stevioside (95.6% purity) at concentrations of 0%, 2.5%, and 5% of the diet of rats for 2 y, resulting in consumption amounts of 0, 970, and 2387 mg ⋅ kg⁻¹ ⋅ d⁻¹ (36). This study evaluated potential effects on physiology (body weight, food consumption, final organ weight), behavior, ophthalmology, biochemistry (blood chemistry, hematology, urine analysis, liver enzymes), and histological changes in tissues. At all of the doses tested, stevioside had no effect on cancer development. No adverse effects were observed in rats fed stevioside at ≤2.5% of diet. At the highest dose (5% of diet), changes were observed for kidney and body weight and survival rates. Therefore, the NOAEL for this study was 2.5% of the diet, or 970 mg ⋅ kg⁻¹ ⋅ d⁻¹, and when converted to steviol equivalents (SEs) was 383 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹.

Applying a 100-fold safety factor to 383 mg SEs results in an ADI of 0–4 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹. The ADI is expressed in SEs because all steviol glycosides are metabolized to steviol, allowing the ADI to apply to all steviol glycosides. Steviol glycosides differ in structure and molecular weight, and therefore contribute relatively different amounts of steviol per gram of steviol glycoside. Therefore, using the conversion factor of 0.33 for Reb A compared with 0.40 for stevioside, which factors in molecular weight and the number of glucose units and steviol per gram, the ADI for Reb A equates to 12 mg ⋅ kg⁻¹ ⋅ d⁻¹ and for stevioside is 10 mg ⋅ kg⁻¹ ⋅ d⁻¹.

An important study that established the safety of steviol glycosides for consumption by pregnant women and children was a reproductive and developmental study of Reb A (>97% purity) (31). Rats were fed ≤2273 mg ⋅ kg⁻¹ ⋅ d⁻¹ of Reb A for 2 generations, while body weight, food intake, growth and development, survival, reproductive performance, and sexual maturation were monitored. No adverse reproductive or developmental effects were observed in any of the generations at the highest dose. Similar results were reported in reproductive toxicology studies with purified stevioside (29, 37). Early studies in rats with crude extracts of S. rebaudiana observed reduced fertility (27) or lower seminal vesicle weights compared with controls (28), but studies with high-purity steviol glycoside extracts (31, 36, 37) did not observe any negative effects on sexual organs, levels of sexual hormones, mating behavior, fertility, gestation length, offspring survival, or sexual maturation. The lack of adverse effects after exposures to high doses of high-purity steviol glycoside before and during critical periods of fertility and pregnancy, during lactation, and throughout growth and development of the offspring to adulthood for 2 generations demonstrates the safety of steviol glycosides for consumption by pregnant women and children at or below the established ADI.

Despite the extensive review and conclusions of safety experts that steviol glycosides are not mutagenic, 2 publications have questioned whether adequate testing of the genotoxic potential of steviol glycosides has been performed (38, 39). In response to their concern, Urban et al. (40) conducted a comprehensive and extensive review of all published in vitro and in vivo studies. Much of the concern was from a few older in vitro studies in which steviol was reported to be mutagenic using a highly specific bacterial strain, Salmonella typhimurium TM677, which requires growth conditions that are not applicable to humans. The review by Urban et al. (40) found consistently negative results for Reb A and steviol, and all negative results for stevioside, with the exception of one study. The in vivo study by Nunes et al. (41), which was positive, has been criticized for its methodology and data interpretation by several reviewers (20, 40, 42, 43). Hence, Urban et al. (40) concluded that the database of in vitro and in vivo studies for steviol glycosides is robust with no evidence that steviol glycosides are genotoxic.

In addition to in vitro and animal studies, human safety studies have also been conducted. Reb A doses of ≤1000 mg/d for 1–4 mo and stevioside doses of 750 mg/d for 3 mo were well tolerated and had no adverse effects on blood pressure or fasting blood glucose in healthy, hypertensive, and type 1 and type 2 diabetic subjects (44–46). Nor were there any significant clinical changes in serum chemistry, hematology, or urine analysis. Most of the safety studies have been conducted on Reb A and stevioside because they are the most abundant steviol glycosides in the S. rebaudiana Bertoni plant. However, all major and minor steviol glycosides are degraded to steviol by human microbiota and therefore share the same metabolic fate. A series of in vitro tests with human fecal homogenates confirmed this for several of the minor steviol glycosides—Reb B, C, D, E, F, and M; dulcoside A; and steviolbioside (12, 23)—thus making the studies on Reb A and stevioside applicable to the minor steviol glycosides as well.

Another concern raised by some is the allergenic potential of steviol glycosides due to the common taxonomy of the stevia plant with plants that can induce hypersensitivity in some individuals (e.g., ragweed, goldenrod, chrysanthemum, echinacea, chamomile, lettuce, sunflower, and chicory). A comprehensive literature search found no evidence of allergenic potential of purified steviol glycosides (47). According to Urban et al. (47), the few cases of allergic reactions that have been reported in the literature occurred before the introduction of high-purity steviol glycosides into the marketplace. Similarly, human studies with high-purity steviol glycosides reported no negative gastrointestinal side effects, such as bloating, gas, diarrhea, nausea, or borborygmus (44–46), which are sometimes associated with certain caloric and nonnutritive sweeteners that include, fructose, sugar alcohols, and allulose (a.k.a. psicose) (48–51).

Overall, the safety data for high-purity steviol glycosides have been thoroughly evaluated and their use as a plant-based, zero-calorie sweetener has been approved across the globe. It has been conclusively determined that foods and beverages containing approved amounts of high-purity stevia leaf extract sweeteners (i.e., steviol glycosides) are safe for all individuals, including children, pregnant and nursing women, and individuals with diabetes.

Dietary exposure

To ensure safety of consumption, the Estimated Daily Intake (EDI) of a food additive should not exceed the ADI. Hence before approval of use, potential intakes are estimated by using proposed food usage levels in various food categories, together with information from food-consumption surveys. The EDI for steviol glycosides has been estimated for various populations (Table 1). In most instances, the EDI for steviol glycosides is less than the ADI, and due to the conservative nature by which they are assessed, estimated intakes are generally recognized as overestimations of what might be actual or average consumer intakes.

Surveys have been utilized in various global jurisdictions to determine daily consumption estimates of high-purity steviol glycosides. The FAO/WHO Joint Expert Committee on Food Additives (JECFA) assessed international dietary exposure estimates with the use of a model that assumed steviol glycosides would replace all sweeteners used in or as food, based on the relative sweetness of steviol glycosides to sucrose (52). The committee estimated maximum intakes of 1.3–5 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹ worldwide. However, the committee acknowledged that these estimates were highly conservative and indicated that actual intakes were more likely to be 20–30% of these values (52). Renwick (53) estimated Reb A intakes for adults, children, and diabetic children using equivalent intake calculations based on existing low- and no-calorie sweetener (LNCS) consumption surveys for North America, Australia, and Europe. For the general population, mean intakes ranged from 0.4 to 0.7 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹, and for adults and children, high intakes (≥90th percentile) were 1.1–1.7 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹.

In 2011, Food Standards Australia and New Zealand (FSANZ), during their review to expand the approval of steviol glycosides, considered 3 dietary exposure assessment models: a 30% market share scenario and 2 “brand loyal” scenarios (54). Although the 90th-percentile dietary exposures of the brand loyal scenarios were 110% of the ADI for Australian children aged 2–6 y and 100% of the ADI for New Zealand children aged 5–14 y, the FSANZ concluded that all 3 models were likely an overestimation. Health Canada (55) used 2 approaches in their exposure assessment in 2012. Method 1 substituted all table-top sweeteners and method 2 assumed maximum authorized use in all food categories. Both approaches resulted in mean intakes that were well below the ADI. Although the maximum use levels (95th percentile) marginally exceeded the ADI for children aged 1–3 and 4–8 y, Health Canada considered these estimates insignificant from a health perspective.

In 2014, following a request from the European Commission, EFSA carried out a revised exposure assessment of steviol glycosides (E 960) to those previously done in 2010 and 2011 (56). The EFSA panel concluded that, overall, the mean exposure estimates remained below the ADI of 4 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹ across all population groups, except for toddlers in 1 country (Netherlands). However, the panel did not consider this to be significant enough to change the outcome of the safety assessment. In a re-evaluation, as part of a US “Generally Recognized As Safe” (GRAS) notice submission (GRN 619) in 2016, estimated intakes of steviol glycosides for the general US population were below the ADI (57). The highest intake was in nondiabetic children, with an intake of 3.28 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹ at the 95th percentile. Dewinter et al. (58) estimated intakes in children with type 1 diabetes who are often at the highest risk of exceeding the ADIs for sweeteners due to their potentially high consumption of sugar substitutes, in their effort to manage a reduced carbohydrate/sugar diet. At the 95th percentile, all age groups had intakes below the ADI, except for 4- to 6-y-olds, who exceeded the ADI at 4.75 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹. Due to the conservative nature of the analyses, the authors concluded that there is little chance that children with type 1 diabetes will exceed the ADIs. To date, based on estimated dietary exposure assessments from different countries and regions of the world, at typical patterns of consumption of foods and beverages containing steviol glycosides, it is unlikely that either adults or children, including diabetic adults and children, will exceed the ADI for steviol glycosides. Although there is no safety concern, it would be valuable to have future research efforts investigate actual dietary intake in adults, children, and subsets of the population who are expected to be high consumers of steviol glycosides and to understand trends over time.

Effect of steviol glycosides on health and related biomarkers

Background

The new WHO sugars guideline recommends that adults and children reduce their intake of added sugars to <10% of total energy intake and recommends a further reduction to <5% for additional health benefits (59). This guideline is part of the WHO’s efforts to halt the increase in diabetes, obesity, and premature deaths by 25% by 2025 (59). The UK Scientific Advisory Commission on Nutrition also recommends a reduction in free sugar to ≤5% (60). For an adult, the 10% and 5% guidelines are equivalent to ∼50 and 25 g sugar/d, respectively. According to WHO estimates, the intake of added sugars among adults ranges from 7–8% of total energy in Hungary and Norway to 16–17% in Spain and the United Kingdom (59). The range for children is higher, varying from 12% in Denmark, Slovenia, and Sweden, to nearly 25% in Portugal (59). In the United States, added-sugar intake has been declining but remains high, with adults and children and adolescents aged 2–18 y consuming 14% and 17% of total energy intake, respectively, in 2011–2012 (61). These amounts are above the recommended maximum of 10% of total energy in the United States (62), as is the case for several other countries.

Postprandial blood glucose and insulin effects

It is well established that the intake of sucrose or glucose creates a postprandial increase in blood glucose and insulin (63). Hence, it is of interest to determine if high-purity steviol glycosides influence postprandial blood glucose and insulin concentrations. A few human studies have examined this effect in single-meal evaluations comparing a reduced-sugar/calorie meal with steviol glycosides compared with a full-sugar/calorie meal, whereas other studies have examined the effect of steviol glycosides in capsules, as supplements, with no dietary manipulation (Table 2). Three randomized controlled trials observed a significant reduction in postprandial blood glucose with purified steviol glycosides utilized in reduced-sugar/calorie meals (64, 65) or in supplement form (66) in healthy subjects and diabetics. Anton et al. (64) observed a significant reduction in postprandial blood glucose (P < 0.01) and insulin (P < 0.05) concentrations when stevia was consumed in a midmorning meal compared to sucrose in lean and obese subjects. Similarly, Jeppesen et al. (65) noted a significant decrease in postprandial blood glucose (P < 0.05), including a 156% lower AUC for blood glucose (P < 0.01) in subjects with type 2 diabetes. Gregersen et al. (66) investigated the postprandial effect of 1000 mg steviol glycosides (91% stevioside) compared to a 1000-mg maize starch placebo given in capsule form along with an isocaloric meal in 12 individuals with type 2 diabetes who had stopped taking hypoglycemic medication before the test. Despite no sugar, carbohydrate, or calorie difference between the test groups, stevioside significantly reduced postprandial blood glucose by 18% (P < 0.004) in addition to the AUC for glucose (P < 0.02) compared to placebo. There was a trend toward an increased insulin response (AUC) and a 40% increase in the insulinogenic index (ratio of AUC insulin to AUC glucose) (P < 0.001) when stevioside was consumed compared with placebo.

Three other studies (20, 67, 68) observed no significant impact on postprandial blood glucose in healthy or diabetic subjects when steviol glycosides were consumed as supplements. However, Jeppesen et al. (67) observed a 45% reduced insulin response in the placebo group (P < 0.05), and an insulin concentration that was maintained in the stevioside group, suggesting that steviol glycosides may have a positive effect on β cell function in subjects with type 2 diabetes. In the intravenous glucose-tolerance test, the insulin response increased after injection of glucose by 21% in the stevioside group compared with placebo (P < 0.05). The patients included in this study may already have been in a late stage of diabetes and therefore may have had limited β cell function, which may explain the different results compared with other human and animal studies.

Overall, when the comparison between steviol glycosides and the control involves a sugar/carbohydrate or calorie differential, postprandial blood glucose reductions have been observed, and this effect is largely due to a sugar and calorie substitution, as observed in the studies by Jeppesen (65), and Anton et al. (64). On the other hand, the postprandial blood glucose decrease observed in the Gregersen et al. (66) study, which had no calorie differential between treatment and control, suggests that, at certain doses, stevioside may have a potential blood glucose–lowering effect in diabetics. These results may not be evident in diabetic subjects who continue taking their hypoglycemic medication, as in the study by Maki et al. (68). Similarly, Maki et al. (68) did not see any change in postprandial insulin concentrations, whereas in studies in which diabetics stopped their hypoglycemic medication, there was evidence of a potential increase in insulin concentrations (66, 67). Additional research is needed to more clearly determine if steviol glycosides have an independent effect on insulin and postprandial blood glucose concentrations in individuals with diabetes, if it is specific to any one steviol glycoside, as well as the mechanism and doses at which these effects may be observed.

Fasting blood glucose and insulin effects

Long-term studies indicate that high-purity steviol glycosides in supplement form within interventions that have no dietary carbohydrate or calorie manipulation do not significantly reduce fasting blood glucose, insulin, or glycated hemoglobin (HbA1c) concentrations (Supplemental Table 1). Studies were conducted in healthy subjects, subjects with type 1 and type 2 diabetes, and hyperlipidemic and hypertensive subjects with a wide range of doses (20, 45, 46, 67, 69–71). These studies had differing protocols involving diabetic subjects, with some continuing their hypoglycemic medications and others stopping just before the start of the study. Although none of the fasting blood glucose measures were significantly changed by the steviol glycoside treatment, it is noteworthy that, in 1 study, 750 mg stevioside/d maintained fasting blood glucose concentrations over a 3-mo period, whereas in the placebo group there was a significant increase compared with baseline among subjects with type 1 diabetes who continued their hypoglycemic medication (46). A similar result was observed in a study by Jeppesen et al. (67), in which 1500 mg stevioside/d was consumed for 3 mo by subjects with type 2 diabetes who had stopped their hypoglycemic medications. A significant difference between treatment and placebo groups for fasting glucose (P < 0.007) and HbA1c (P < 0.01) was observed. These findings suggest that stevioside at levels above the ADI may help maintain a static diabetic state, which could be beneficial to individuals with diabetes in minimizing or slowing down the progression of diabetes. Furthermore, a meta-analysis of several of these studies by Onakpoya and Heneghan (72) showed a small but significant reduction in fasting blood glucose (–0.63 mmol/L; P < 0.00001). However, the clinical relevance of a reduction of 0.63 mmol/L observed in the meta-analysis may be limited.

Jeppesen and Larsen (73) also examined the effect of supplementing 500 mg steviol glycosides, together with postexercise oral carbohydrate, compared with isocaloric carbohydrate supplementation on muscle glycogen resynthesis in 15 male cyclists. The glycogen resynthesis rate was increased by 35% (P < 0.02) and glycogen concentrations were significantly higher (P < 0.009) with steviol glycosides compared with placebo. More research is needed to understand how steviol glycosides may confer these effects.

Potential mechanisms related to blood glucose

It is clear that one indirect way in which steviol glycosides and other LNCSs lower postprandial blood glucose concentrations is through the displacement of sucrose or other carbohydrates (74). However, for steviol glycosides, a few in vitro and animal studies suggest a potential independent and more direct mechanism involving insulin secretion, signaling, and release; upregulation of key genes; and enhanced glucose absorption in primarily diabetic models. Jeppesen et al. (75) was the first, to our knowledge, to show that both stevioside and steviol (1 nmol/L to 1 mmol/L) dose-dependently enhance insulin secretion from incubated mouse islets in the presence of glucose (P < 0.05). The insulinotropic effects of stevioside and steviol were critically dependent on the glucose concentration and occurred at ≥8.3 mmol glucose/L (P < 0.05). To determine if stevioside and steviol act directly on pancreatic β cells, the β cell line INS-1 was used. Both stevioside and steviol potentiated insulin secretion from INS-1 cells (P < 0.05).

Animal studies of steviol glycosides suggest an effect on insulin secretion and sensitivity and gluconeogenesis. Jeppesen et al. (76) performed an intravenous glucose-tolerance test with and without 0.2 g stevioside ⋅ kg⁻¹ ⋅ d⁻¹ in type 2 diabetic Goto-Kakizaki and normal Wistar rats. In diabetic rats, stevioside significantly suppressed the blood glucose response (incremental AUC, P < 0.05), while concurrently increasing the insulin response (incremental AUC, P < 0.05). Chen et al. (77) reported that 0.5 mg stevioside ⋅ kg⁻¹ ⋅ d⁻¹ provided by gastrointestinal gavage lowered blood glucose concentrations in normal rats, as well as in 2 models of diabetic rats in a dose-dependent manner, not only by enhancing insulin secretion but also by slowing down gluconeogenesis in the liver by decreasing concentrations of phosphoenol pyruvate carboxykinase, an enzyme involved in the metabolic pathway of gluconeogenesis. Nordentoft et al. (78) in a 9-wk intervention study in genetically obese diabetic KKAy mice treated with 20 mg ⋅ kg⁻¹ ⋅ d⁻¹ observed that the stevioside derivate isosteviol had a high bioavailability from the colon and improved glucose and insulin sensitivity by upregulating the gene expression of key insulin-regulating genes and insulin transcription factors. Chang et al. (79) observed that a single oral administration of 0.5 mg stevioside ⋅ kg⁻¹ ⋅ d⁻¹ for 90 min decreased plasma glucose concentrations and reversed the glucose-insulin index, a measure of insulin action on glucose disposal in rats fed fructose-rich chow for 4 wk. Repeated administration of stevioside delayed the development of insulin resistance in these rats and increased the response to exogenous insulin in streptozotocin-diabetic rats. Philippaert et al. (80) showed that 0.5 mg stevioside ⋅ kg⁻¹ ⋅ d⁻¹ given orally 2 h before a glucose-tolerance test significantly lowered blood glucose concentrations in normal wild-type mice but not in the transient receptor potential cation channel subfamily M member 5 (TRPM5) mice. TRPM5 is a Ca²⁺-dependent cation channel found in type II taste receptor cells on the tongue and in insulin-producing β cells in the pancreas. TRPM5 knockout mice have decreased glucose tolerance due to impaired glucose-induced insulin release.

A study of Reb A on metabolic syndrome outcomes suggests similar outcomes to stevioside. Jeppesen (81) fed rats a high-fructose diet for 16 wk followed by 8.4 mg Reb A/d and 16.8 mg aspartame or high-fructose corn syrup (HFCS)/d at 13% of total caloric intake for 8 wk. Incremental AUC glucose was significantly lower for the Reb A group compared with the HFCS group (P < 0.05) after a glucose-tolerance test. Insulin resistance measured by HOMA-IR (P < 0.005) and hepatic TG content (P < 0.05) were significantly reduced in the Reb A and aspartame groups. In addition, the expression of fatty acid metabolism genes Srebf1 in liver and Fas in liver and muscle were significantly lower in the Reb A group compared with the HFCS group (P < 0.001).

Overall, the research supports a beneficial effect and no adverse effects of steviol glycosides for blood glucose management when steviol glycosides are used to reduce or substitute sugar and calories in a food, meal, or diet. The longer-term safety studies that range from 3 mo to 1 y in normal individuals and those with diabetes indicate that steviol glycosides are safe and have a neutral effect on fasting blood glucose, insulin, and HbA1c at doses of ≤1500 mg/d. One meta-analysis suggested a modest reduction in fasting blood glucose. The doses studied in several long-term studies were well above the ADI. Some preclinical and clinical studies suggest a potential independent effect of steviol glycosides in lowering postprandial blood glucose concentrations, enhancing insulin secretion, and improving insulin sensitivity in subjects with diabetes, with some mechanistic evidence for these effects. Additional clinical studies are needed to clarify and confirm these findings.

Energy intake and weight control

Full replacement of caloric sweeteners with LNCSs in foods and beverages can provide a desirable sweet taste with little or no sugar and calories. In light of several recent policy recommendations to reduce sugar in the diet (59, 62, 82), LCNSs, including steviol glycosides, offer a simple and effective way to reduce both sugar and calories in the diet and thereby also offer a helpful way to manage both energy intake and body weight.

Steviol glycosides

To date, 2 studies (64, 83) have evaluated the effect of steviol glycosides on satiety and energy intake (Table 3). Anton et al. (64) observed no increase in subjective satiety but found that energy intake was significantly decreased over the day when 2 reduced-energy/sucrose preload meals with steviol glycosides were consumed 20 min before an ad libitum lunch and dinner. Thirty-one subjects consumed 309 kcal less during the steviol glycoside compared with the sucrose treatment (P < 0.001). There were no differences in energy intake at lunch or dinner; therefore, the daily energy difference was primarily due to the energy difference in the 2 preloads. Energy compensation was 24% during the steviol glycoside period. A second study evaluated the effects of steviol glycosides consumed in water compared with a sucrose control 1 h before an ad libitum lunch in 30 men and observed no difference in satiety ratings but noted a total daily energy intake reduction of 70 kcal (83). The energy compensation during the steviol glycoside period was 73%. The higher energy compensation in this study than in the first could possibly be attributed to several factors, including the number and use of different preloads, the time interval between the preload and the ad libitum meal, and the fact that the study by Tey et al. (83) was not statistically powered to assess energy intake differences but was powered to detect a 30% difference of the blood glucose treatment. Across the 2 studies the average energy compensation was ∼50%, similar to the average energy compensation observed for other LNCSs (84).

LCNSs

Due to the absence of clinical trials on the effect of steviol glycosides on body weight, the symposium included a brief review of the impact of LNCSs on energy intake and body weight, because it would be anticipated that the effect would be similar for steviol glycosides if a study were carried out. Research shows that there is no precise physiologic balancing of energy intake against energy expenditure. The consumption of energy either in excess or deficit of immediate energy requirements is not fully compensated for by adjustments in intake at the next meal or at subsequent meals (85). Hence, reduced energy intake by LNCSs use should be helpful to those attempting to maintain or lose weight. Consistent with this, a recent meta-analyses of 69 acute and long-term randomized controlled studies in human participants between 1970 and 2015 found clear evidence that the consumption of LNCSs in place of (some) sugar in the diet reduces energy intake and body weight (84). Despite these findings, claims persist that LNCSs hinder rather than help appetite and weight control.

Based on a rodent model, one claim has suggested that by “decoupling” sweetness from caloric content, LNCSs disrupt the animal's learned ability to regulate energy intake (86, 87). In these studies, rats that consumed saccharin-sweetened yogurt increased their intake of food, which led to increased weight gain and body fat accumulation and decreased caloric compensation compared with rats that consumed glucose-sweetened yogurt (86, 87). A basic premise underlying these studies is that sweet taste is a valid predictor of increased energy intake. However, this can be challenged, because sweetness does not reliably predict the energy content of foods (88). Furthermore, there is also the question of whether rats, or humans, rely only on simple taste-nutrient relations to control energy intake. It is more likely that signals triggered by nutrients detected in the gut postabsorptively dominate in influencing satiety (85). Recent research has failed to replicate the earlier “decoupling” findings. In 2 experiments, Boakes et al. (89, 90) observed that rats intermittently fed glucose gained more weight, fat mass, or both than did rats intermittently fed saccharin. This is opposite to the results reported by Swithers et al. (86). The discrepancy between these 2 sets of results appears to be explained by the fact that the study by Swithers et al. (86) excluded rats that showed low acceptance of the saccharin-sweetened yogurt. Boakes et al. (90) showed that this biases the sample toward faster-growing rats, because saccharin acceptance is associated with later weight gain with chow feeding. In other words, the result reported by Swithers et al. (86), and quoted widely to support the LNCSs “confuse your body” claim, is a procedural artefact. The results from Boakes et al. (89), on the other hand, are plausibly explained by a lack of full compensation for the higher energy content of the glucose-sweetened yogurt. This was confirmed in a systematic review in which 59 of 68 animal studies of continuous exposure to LNCSs showed no significant weight change or decreased body weight (84).

Another claim suggests that repeated exposure to sweetness encourages a “sweet tooth” and therefore the increased intake of sweet, energy-containing foods and drinks (91, 92). This assertion was tested in 2 recent studies. In a sample of 39 participants, the desire to consume apple juice, apple, and apple pie was significantly reduced (P < 0.05) when an LNCS drink was consumed before the meal than when water was consumed (93). A second study tested the effect of consuming sweet drinks on sweet and savory food intake. On 3 separate occasions, 50 participants were presented with a savory snack (Doritos, Frito-Lay) and a sweet snack (chocolate chip cookies) after the consumption of water, LNCS soda, or a regular sweetened soda (93). The consumption of the sweet snack was significantly reduced after the intake of the LNCS soda (P < 0.05) and the regular soda (P < 0.01) compared with water. In contrast, the intake of the savory snack was not significantly affected by the ingestion of the sweetened beverages. These results are consistent with the phenomenon of “sensory-specific satiety,” which is the reduction in liking or reward value of a recently eaten food compared with a recently uneaten food or taste (94, 95). It is also consistent with the findings from a 6-mo intervention study in which participants who substituted caloric beverages with LNCS beverages significantly reduced their intake of desserts compared with participants who substituted caloric beverages with water (96). In another study, participants who reduced their intake of sweet foods and drinks for 3 mo showed an increase in perceived sweet-taste intensity (at low concentrations of sucrose), but no change in perceived pleasantness of sweet, test products (97). Finally, randomized controlled trials have generally found no effect on body weight between a diet moderately high in sugars and a diet in which free sugars were replaced by the isoenergetic exchange of lower-sugar carbohydrates (98), again showing that sweetness per se does not encourage increased energy intake.

For LNCSs to successfully contribute to reduced energy intake, it is necessary that compensatory energy intake not occur. To address this issue, a systematic review and meta-analysis examined both short-term (≤1 d) and sustained (>1 d) randomized controlled studies (84). The short-term analysis evaluated 218 comparisons from 56 articles that examined the effect of a LNCS preload compared with sugar, unsweetened product, water, nothing, or placebo capsules on subsequent energy intake. Most of the comparisons (83%) were LNCSs compared with sugar, in which it was observed that LNCSs, when substituted for sugar, consistently reduced short-term energy intake. LNCS intake compared with sugar resulted in 70% energy compensation in children and 43% compensation in adults, leading to an average compensation across all studies of 50%. Energy intake also did not differ for LNCS comparisons with water, unsweetened product, or nothing. The sustained energy intake analysis included 10 comparisons from 9 studies that ranged from 10 d to 1 y in overweight, obese, and normal-weight participants; and in all instances, the use of LNCSs led to a reduction in energy intake. Results of another study completed after this review were consistent with the findings of Rogers et al. (84), in which it was noted that LNCS beverage consumption with meals did not increase total energy intake, macronutrient intake, or sweet foods selected, either in those who were habitual or nonhabitual consumers (99), contrary to the concern that LNCSs might increase energy intake by decoupling sweetness with energy content, or by enhancing preference for sweets, or other potential mechanisms reviewed by Mattes and Popkin (100).

The relation between LNCS intake and body weight has been examined by several observational (i.e., prospective cohort) studies and randomized controlled trials. Randomized controlled studies provide the highest quality of evidence. Table 4 summarizes the findings of systematic reviews and meta-analyses (74, 84, 101–106). Results from 7 systematic reviews of prospective cohort studies were mixed, with the majority showing no clear trend. One meta-analysis observed a very slight decrease in BMI (kg/m²; –0.002) (84), whereas another observed a slight increase in BMI (kg/m²; 0.03) and no significant association with body weight or fat mass (102). In observational studies, it is not possible to control for all potential confounding factors and therefore the possibility of residual confounding remains, as well as the possibility of reverse causality (106). Of the 6 systematic reviews and 2 meta-analyses of randomized controlled trials, most showed a decrease in body weight or BMI with LNCS use. Both meta-analyses reported that LNCS use was found to reduce BMI, body weight, or both (84, 102). Miller and Perez (102) found that LNCS use was significantly associated with reduced body weight (–0.80 kg), BMI ( kg/m²; –0.24), waist circumference (–0.83 cm), and fat mass (–1.10 kg). Similarly, Rogers et al. (84) reported a significant reduction in body weight when LNCSs were substituted for sugar (–1.35 kg) or water (–1.24 kg).

Collectively, the research to date shows that the consumption of LNCSs, including steviol glycosides, consistently help reduce energy intake, contrary to the suggestion that LNCSs might increase energy intake. In addition, studies show that exposure to sweetness does not train taste preference and encourage a “sweet tooth.” There is, in fact, no human clinical study that would suggest that a sustained exposure to “sweetness” with LNCSs would lead to an increase in energy intake. With regard to steviol glycosides, despite differences in study design, the 2 available studies (64, 83) showed an energy-reduction benefit, with an average energy compensation of 50%. Overall, the current evidence is consistent with a current expert consensus article (107), which concluded that LNCSs help to reduce energy when used in place of higher-energy ingredients. Claims that LNCSs increase appetite and body weight are clearly contradicted by evidence showing that the consumption of LNCSs can be expected to contribute to healthy weight management. It is also safe to assume that steviol glycosides would likely result in similar weight-reduction benefits observed in randomized controlled studies of other LNCSs.

Blood pressure

Six randomized clinical trials with 8 clinical study arms have investigated the effect of steviol glycosides on blood pressure from 4 wk to 2 y. Two clinical arms conducted in healthy adults with normal blood pressure observed no significant differences between the consumption of 750–1500 mg/d steviol glycosides and the placebo control (44, 46). Four clinical arms found no significant impact of steviol glycosides on blood pressure in individuals with type 1 and type 2 diabetes, but in all 4 instances, the subjects continued taking their blood pressure medications if they were hypertensive (45, 46, 67). Subjects with mild to moderate hypertension who were not taking blood pressure medication were investigated in 2 studies, and both showed a modest blood pressure–lowering effect with 750–1500 mg stevioside/d (70, 71). The steviol glycoside interventions were provided in supplement form with no dietary manipulation, with the purpose of examining their safety and independent effect on blood pressure.

A meta-analysis of 7 randomized controlled trials that assessed steviol glycosides in both acute single-meal and long-term settings showed a nonsignificant difference in systolic blood pressure but a significant decrease in diastolic blood pressure (–2.24 mm Hg; P = 0.03) (72). However, significant heterogeneity was observed, likely due to differences in the composition of the steviol glycosides, doses utilized, continued use of blood pressure and antidiabetic medications by subjects, and the inclusion of subjects with normal blood pressure. Most of these studies were designed to investigate the safety of steviol glycosides within these contexts, with several studies using doses that were 3–4 times the ADI with no negative impact, further supporting the safety of steviol glycosides.

Gut microbiota

The human gut microbiota is a large and complex population of microorganisms. More than 1000 species have been identified in total, with ∼160 being present in the gut of any one individual (108). More than 90% of the species fall into 2 main phyla, Firmicutes and Bacteroidetes; other common phyla include Actinobacteria, Proteobacteria, Verrucomicrobia, and Fusobacteria (109). There is also evidence that the microbiota may also be involved in obesity and type 2 diabetes (110). It has, however, proven to be more difficult to identify the microorganisms involved in these conditions.

The relative proportions of the phyla and their component genera and species, as well as gut microbial metabolism, can vary markedly between individuals and can be influenced by a variety of factors, including early colonization in the immediate postnatal period, host genetics, and exposure to drugs and environmental chemicals (111). Mounting evidence, however, indicates that diet, both habitual and long-term and shorter-term dietary changes, appears to be the most significant factor influencing the overall composition of the gut microbiota and its functionality.

Because of their extensive use in foods, the interactions of LNCSs and the gut microbiota have been the subject of numerous studies in laboratory animals and human subjects, although LNCSs are unlikely to have a clinically meaningful impact because they are consumed at such low amounts. Nevertheless, some studies on saccharin, aspartame, and sucralose have shown effects on microbiota composition or metabolism, but only at very high doses above normal human consumption, or in studies with design issues or lacking appropriate controls (112–116). LNCSs are a structurally diverse group of compounds that have very different metabolic fates after consumption, as reviewed by Magnuson et al. (15). Most (e.g., acesulfame K, saccharin, aspartame, and sucralose) are not metabolized by gut bacteria. The only 2 exceptions are steviol glycosides and cyclamate. The latter is converted by microbiota to cyclohexylamine, which is subsequently absorbed and excreted in urine (117).

Studies on the impact of steviol glycosides on the gut microbiota are few. Gardana et al. (17) incubated human fecal suspensions with stevioside or Reb A for 24 h. Decreases were seen in numbers of total anaerobes, Bacteroides, and Lactobacilli with stevioside, and in total aerobes, Bifidobacteria and Enterococci in incubations with Reb A. In all cases, the changes in number were small (<1 log). Similarly, Kunová et al. (118) noted in another in vitro study that the growth of lactobacilli and bifidobacteria strains was poor in the presence of steviol glycosides compared with a glucose control. Denina et al. (119) also observed the lack of growth of Lactobacillus reuteri strains after the incubation of stevioside and Reb A for 24 h. A study in BALB/c mice given Reb A orally for 4 wk at 5.5 mg or 139 mg ⋅ kg⁻¹ ⋅ d⁻¹ (1.8 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹ or 46 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹) compared with water reported no changes in viable counts of the major groups in feces, or in diversity indexes of total bacteria (120). The only difference was an increased diversity of Lactobacilli at the higher dose, which was >10 times the ADI of 4 mg SEs ⋅ kg⁻¹ ⋅ d⁻¹. Thus, the current evidence indicates that steviol glycosides have minimal impact on the gut microbiota.

Although there is no effect of steviol glycosides on the gut microbiota, data do indicate that steviol glycosides are metabolized by gut bacteria. The microbiota provides an important role in the breakdown of dietary ingredients by providing enzymes that are not present in humans (121). Although glycosylases are common among members of the microbiota, Gardana et al. (17) found that the ability to deglycosylate steviol glycosides appears to reside only within the Bacteroides genus. Cultures of Clostridia, Bifidobacteria, coliforms, Lactobacilli, and Enterococci tested were unable to metabolize stevioside or Reb A. Human variability in the hydrolysis of steviol glycosides is expected to be minimal because Bacteroides is by far one of the most abundant bacterial groups found in the large intestine (122).

Dental caries

The relation between the consumption of sugar and the incidence of dental caries has been well established. Two short-term clinical studies were conducted with stevia. Brambilla et al. (123) showed that the plaque pH of sucrose (P < 0.01) was significantly lower after a single rinse compared with stevioside or Reb A at identical concentrations at 5, 10, 15, and 30 min after rinsing in 20 adults. The reduced growth of Streptococcus mutans in a biofilm model was also observed with stevioside and Reb A. Zanela et al. (124) reported that the accumulation of plaque in 200 children was not reduced in daily mouth rinses containing 0.5% stevioside with 0.05% sodium fluoride compared with 0.12% chlorhexidine with 0.05% sodium fluoride. Counts of S. mutans did not differ between the groups, but the results may have been confounded because 20% of the children in all groups had low levels of S. mutans at baseline. Furthermore, a comparison of stevioside with sucrose may have been a more appropriate comparison rather than chlorhexidine. A study in rat pups infected with Streptococcus sobrins observed that after 5 wk of treatment, stevioside and Reb A were noncariogenic, in contrast to sucrose for which deep fissure and surface caries and the highest number of S. sobrin counts were noted (125). Two additional in vitro studies report on the effects of stevia compared with typical pharmacologic interventions. In one study the inhibitory effect of chlorhexidine was greater against S. mutans growth than stevia extract in aqueous and alcoholic solutions (126), and another study showed positive but lower antimicrobial properties of stevia extracts compared with 2 positive controls, vancomycin and azithromycin (127). Overall, the data suggest that steviol glycosides are not cariogenic and may have beneficial effects in preventing dental caries compared with nutritive sweeteners (e.g., sucrose, HFCS, etc.). However, additional long-term human studies with the use of stevia in place of cariogenic nutritive sweeteners are warranted.

Naturality and processing of steviol glycosides

High-purity stevia is extracted and purified from stevia leaves in a manner that is similar to that of sucrose from sugar cane. Specific variables involved in the extraction and purification of steviol glycosides can vary among stevia producers, but in all instances, the process starts with the leaves of the S. rebaudiana Bertoni plant, which are harvested, dried, and crushed (128, 129). They are then steeped in warm water, similar to a tea infusion (130). Steviol glycosides are soluble in water due to their monosaccharide moieties and can be extracted in large-scale commercial processes with a yield of ≤100%. This water extract is dark brown because of other constituents in the leaves, such as protein, fiber, dyes, polyphenols, minerals, and salts, which are also extracted. Purification steps remove the nonsugar constituents, and the remaining steviol glycosides are spray-dried to an off-white intermediate that contains 80–95% steviol glycosides (131). This end product is further purified by crystallization using water or ethanol mixtures to a white end product with a purity of ≥95%. These purification steps are physical processes used to remove unwanted constituents of the leaves, which enable steviol glycosides to be concentrated (13). The process of extraction and purification does not affect the chemical identity of the steviol glycosides, allowing them to remain as they were when located intact in the leaves. Some have called into question this conclusion and therefore the naturality or natural authenticity of high-purity stevia leaf extract. To address this question, a recent study determined whether steviol glycoside molecules are altered or if their pattern is changed during the process of extraction and purification from the leaves of the stevia plant to the high-purity end product (131).

Three separate batches of a large-scale commercial extraction and purification process, which included the dried leaves (SL), the first water extract (ESL), and the final product, a stevia leaf extract with a purity of >95% (SLE95), were examined (131). All 9 steviol glycosides (Reb A, B, C, D, and F; rubusoside; steviolbioside; dulcoside A; stevioside) listed in the JECFA's 2010 specification (129) were detected and were well separated by using HPLC and MS detection. The samples from all 3 processing steps showed comparable chromatograms with the same pattern and retention times per the US Pharmacopeia reference standard, with the exception of Reb D, which eluted quite early and could only be detected in the end product. An MS detector was applied, with HPLC conditions that were comparable to those applied in the first round of testing, and the identities of all 9 steviol glycosides including Reb D were confirmed unambiguously in the leaves, the first water extract, and the high-purity end product (131).

The relative distribution of the sweeteners for each batch was also calculated. It was found that the relative amounts of Reb A, C, and F; dulcoside A; and stevioside were comparable across samples of SLE95, ESL, and SL. A slight tendency toward depletion was seen for rubusoside, Reb B, and steviolbioside in the SLE95 samples in comparison to the ESL and SL samples in each series. However, the most salient point is that the 9 steviol glycosides detected in the leaves were found in the water infusion (ESL samples) and the high-purity end-product powder (SLE95 samples) in a similar pattern. These results confirm that steviol glycosides tested in this study were not chemically modified or degraded during the traditional large-scale commercial extraction and purification processes used to produce high-purity steviol glycoside sweeteners, thus providing support for the natural authenticity of steviol glycosides.

Alternate technologies for steviol glycoside production

Innovations in the production of “steviol glycosides” by glycosylation, biotransformation (also known as bioconversion), and from genetically modified yeast have focused on reducing cost and improving taste by minimizing the lingering bitter aftertaste or off flavors that have been found with some steviol glycosides.

Glycosylation is based on the premise that taste is improved when ≥1 sugar moieties (usually glucose units) are added to the steviol glycoside molecules extracted from the stevia plant (132, 133). The process starts with purified stevia leaf extract that is produced using traditional extraction and purification methods. The extract is then treated with the enzyme cyclodextrin glycosyl transferase, which enables the transfer of glucose from a sugar source, such as corn starch, to steviol glycosides, thus modifying their chemical structure. The end product of glycosylation is a structurally modified form of stevia that consists of several new glycosylated steviol glycosides that are not found in the stevia plant, and with less of the unaltered steviol glycosides.

The recent discovery of the genes that encode the biosynthesis of steviol glycosides such as Reb A, D, and M has led to the development of Reb A, D, and M production in genetically modified yeast strains of Saccharomyces cerevisiae (134, 135) and Yarrowia lipolytica (136). These strains of yeast are genetically engineered to express the steviol glycoside metabolic pathway of the stevia plant, allowing them to produce the enzymes, the intermediates, and steviol glycosides such as Reb A, D, and M in a fermenter with corn dextrose or glucose as a sugar source. Steviol glycosides produced from genetically modified yeast are not derived from the stevia plant and do not use any part of the stevia plant in the process.

Another recently developed technology, known as biotransformation or bioconversion, starts with traditionally extracted steviol glycosides such as stevioside or Reb A that are then transformed with the use of multiple genetically modified yeast—such as, Pichia pastoris strains A and B, as noted in a 2017 US GRAS notification (137). These genetically modified yeast are engineered to contain specific enzymes of the biosynthesis pathway of steviol glycosides that selectively transfer glucose units from a glucose source, such as corn dextrose, to the extracted steviol glycoside, for e.g., stevioside, converting it to Reb E and then to Reb M or other desired steviol glycosides. The end products of biotransformation are identical to steviol glycosides found in the plant. It starts with an extract from the plant that is then transformed.

Traditional extraction and purification of steviol glycosides from the stevia leaves remain a good way to produce high-purity steviol glycosides that are not genetically modified and do not affect the natural authenticity of the product. Recent proprietary traditional nongenetically modified organism (non-GMO) breeding methods have resulted in new stevia varieties, such as a variety known as Starleaf by PureCircle Ltd., that has been developed to contain the desirable steviol glycosides, Reb M and D, at levels that are 20 times higher than historically known in stevia plant varieties (138). These breeding methods are making available better-tasting steviol glycoside sweeteners that are plant-based, enabling greater reductions in the sugar content of foods and beverages.

Taste and sensory aspects

The intensity of sweetness and flavor profiles differ widely among the different steviol glycosides (Supplemental Table 2). In general, the sweetness potency of LNCSs, including steviol glycosides, is dependent on sucrose reference concentrations. For example, the relative sweetness of Reb A and stevioside is 180–350 times than that of sucrose in a 2.5–10% aqueous solution. Advances in stevia research have found that some of the minor steviol glycosides, such as Reb M and D, have a higher sweetness intensity, are more sugar-like in taste, and have minimal aftertaste compared with steviol glycosides such as Reb A and stevioside [(139–142); PureCircle, 2017 unpublished data]. The relative sweetness of all of the minor steviol glycosides compared with that of sucrose is not fully known, because the focus has been on combinations of steviol glycosides. However, from research on proprietary combinations, it is known that the minor steviol glycosides contribute to both sweetness and flavor modification, which can influence how a combination works in a given food or beverage matrix compared with another (PureCircle, 2017 unpublished data).

Replacing sugar in food and beverage products is not simple because sugar provides texture, viscosity, and mouthfeel and has no lingering aftertaste, which not all LNCSs can mimic perfectly. For example, in baking, sugar not only provides sweetness, it also contributes to crispness, cell structure, browning, tenderization, and shelf stability, all of which influence mouthfeel, sweetness, flavor perception, and control of water activity. Therefore, when sugar is reduced in a baked food, bulking agents such as maltodextrin, sugar alcohols or fibers, and hydrocolloids or proteins are used with stevia to mimic the characteristics of sugar and to provide moisture and texture that full-sugar versions provide. In recent studies, for 20%- to 50%-reduced-sugar muffins with stevia, cocoa fiber and inulin were used to provide the optimal level for texture, sweet taste, and flavor (143, 144). Stevia is generally heat stable and may even enhance flavors in baked goods, such as salt, spice, and brown aromatics (PureCircle, unpublished data).

Commercially sold, high-purity stevia leaf extracts may contain either a single steviol glycoside (e.g., Reb A) or various combinations of steviol glycosides. Unlike other sweeteners, stevia's sweetness is naturally derived from >40 steviol glycosides, which makes stevia more complex to work with than single-compound sweeteners. In addition, some of the challenges of LNCSs, including stevia, are that they can have “off” tastes, such as bitter and metallic, slow-onset, and sweet tastes that linger (145). Reb D and Reb M have a relatively clean sweet taste, whereas stevioside and Reb A, although sweet, can also impart bitter, metallic, and or licorice-like tastes to varying degrees depending on the amount used (5). Aside from the range of sweetening potency, each of the steviol glycosides have different solubilities and exhibit unique sensory and functional attributes that also allow them to modify or enhance flavors, such as lemon, fruity, floral, brown, and spicy notes.

Most consumers do not want to compromise on taste and prefer the taste of sucrose. Therefore, the goal when working with high-potency LNCSs is to replicate as closely as possible the taste and functionality of sucrose. Taste perception is influenced by product matrix and, in the case of stevia, sweet taste can be significantly improved through the use of unique high-purity steviol glycoside combinations, optimally designed for a given food or beverage matrix. These innovations point to taste advantages that are far superior than the use of any single steviol glycoside, such as Reb A or Reb M, alone (146), thus helping to achieve maximum sugar reduction while imparting a more sugar-like taste without adding calories or bitter off notes. Figure 2 shows results from a sensory study in 30 panelists who compared a sucrose control with 2 high-purity stevia leaf extract products in acidified water—namely, Reb A (97%) and a proprietary ingredient that contained a combination of steviol glycosides (PSB-1198) sold by PureCircle Ltd. Acidified water was used because it is representative of characteristics of select market beverages that use stevia. Panelists reported a lingering off taste and less upfront sweetness for the Reb A compared with the PSB-1198, showing the advantage of this steviol glycoside combination. The results indicate that the taste profile of PSB-1198 was closer to the taste profile of sucrose (146).

Research in the area of taste science can offer additional clues to enhancing stevia's overall palatability. Humans perceive 5 basic tastes: sweet, umami, bitter, salty, and sour. Of these, sweet and bitter tastes are of the most relevance to stevia (147). Taste perception can change when multiple taste stimuli are presented together in a food or beverage compared with 1 stimulus, known as a binary taste interaction (148). The sweet and bitter tastes found in steviol glycosides interact and the overall bitterness threshold of steviol glycosides may be affected (149). Sweet and bitter tastes are detected by different taste receptor cells (147, 150). According to Bachmanov et al. (147), human taste perception, especially bitter tastes, can vary greatly among individuals due to genetic variation. A sensory study in 10 trained panelists combined with in vitro cell-based receptor assays determined how steviol glycosides are sensed by the tongue (149). Results indicated that 2 receptors, TAS2R4 and TAS2R14, mediate the bitter taste in steviol glycosides. The researchers also noted that there are 3 key structural features that appear to modulate the sweet and bitter taste in steviol glycosides—namely, glycone chain length, pyranose substitution, and the C16 double bond. Steviol glycosides that had more glucose molecules attached to them were sweeter and less bitter.

Research on sweet taste receptor cells may also be utilized to optimize the taste of steviol glycosides. The area of a taste receptor cell that tastants bind to is referred to as a docking site (151). Findings from a docking study on 8 steviol glycosides showed significant variation in the docking positions of all steviol glycosides tested. Docking scores predicted the sweetness potency of steviol glycosides. The researchers noted that the interaction of the C13 and C19 glucose molecules with a specific set of active docking sites was responsible for their characteristic taste (152). These results suggest that modifying steviol structures and enabling their binding toward a specific point in the sweet taste receptor cells may be a useful means of enhancing the taste quality and sweetness index of steviol glycosides.

Regulatory status

The safety and use of steviol glycosides has been reviewed and considered by multiple scientific bodies and regulatory agencies around the world. High-purity stevia leaf extracts have been approved or adopted for use in foods and beverages in >150 countries and regions, including the United States, the European Union, the Middle East, Australia, New Zealand, Canada, China, Japan, Korea, Malaysia, India, Mexico, Brazil, Chile, Paraguay, Argentina, Egypt, Ghana, South Africa, Kenya, and many other countries in Asia, Europe, Latin America, and Africa.

In the United States, extracts from stevia have been used as dietary supplements since the 1990s (18), and the use of high-purity steviol glycosides in foods and beverages has been determined to be GRAS on the basis of the evidence from published toxicology studies and the review of product-specific data by qualified experts who evaluate safety of use (153). High-purity Reb A received GRAS status (GRN 252) with a “no objection” letter from the US FDA in 2008 (130). To date, according to the US FDA's GRAS Notice Inventory, the agency has issued >40 “no objection” letters on GRAS notices for steviol glycosides. A high-purity stevia specification consisting of 9 steviol glycosides (Reb A, B, C, D, and F; rubusoside; steviolbioside; dulcoside A; stevioside) at a minimum 95% purity was established by the Codex Alimentarius Committee in 2010 (129). In 2011, Codex adopted steviol glycosides as a food additive with the establishment of food-use standards across a variety of food and beverage categories. The French Food Safety Authority was the first in Europe to assess the safety of Reb A and approve its use in 2009. A favorable scientific opinion by EFSA (14) led to the approval of 10 steviol glycosides by the European Commission in 2011, which included the 9 approved by JECFA and Reb E. After an initial approval in 2008, FSANZ made revisions in 2010 and 2011 to include higher levels of use and select food categories. Hong Kong and Swiss approvals occurred in 2010, and between 2011 and 2012, Health Canada and several countries in Asia, Latin America, and the Russian Federation approved the use of steviol glycosides for foods and beverages. Between 2014 and 2016, high-purity steviol glycosides were approved in India, several Southeast Asian countries, and the Gulf Cooperation Council countries of the Middle East.

Investigations with lower-purity products, such as RebA-80 (80% steviol glycoside purity) and RebA-50 (50% steviol glycoside purity), compared with pure Reb A led to the realization that mixtures of steviol glycosides may offer superior taste to that of pure Reb A. This resulted in the development of several stevia sweetener products composed of different combinations and purity levels. In addition, the study of minor steviol glycosides led to an improved understanding of their taste and functionality. As a result, between 2013 and 2016, there have been 3 US GRAS notices that include Reb M, Reb D, or both (134, 154, 155). GRN 473 and 512 are for Reb M extracted from the leaves of the stevia plant (154, 155), whereas GRN 626 is for Reb M and D produced by a genetically engineered strain of yeast, S. cerevisiae (134). Reb M has also been approved by EFSA, FSANZ, and Health Canada. A recent GRAS notice (GRN 619) with a no-objection letter from the US FDA in 2016 expands the use of stevia to include the safe use of ≥40 steviol glycosides (57). In addition, JECFA's 2017 safety review and proposal supersedes previous specifications, by proposing the use of all natural-origin steviol glycosides (≥50) containing a steviol backbone conjugated to any number or combination of the principal sugar moieties in any of the orientations occurring in the leaves of S. rebaudiana Bertoni, including glucose, rhamnose, xylose, fructose, and deoxyglucose (156). This new proposed specification is expected to be adopted by Codex in the 2018.

Of the 2 known genetically modified yeast Y. lipolytica (136) and S. cerevisiae (135) engineered to produce steviol glycosides, to date, JECFA has approved the use of Reb A produced “from multiple gene donors expressed in Yarrowia lipolytica” at a minimum of 95% purity (157). Additional ingredients that use alternate technologies have been approved or have GRAS status. Between 2011 and 2016, several US GRAS notices with no-objection letters from the US FDA (e.g., GRN 452, 656, 448, 375, 337, and 607) for glucosylated steviol glycosides allowed their commercialization (132, 158–162). China, the United States, Japan, Malaysia, and Korea also allow the use of glucosylated stevia ingredients. In addition, 2 steviol glycoside ingredients (GRN 667 and 715) produced via bioconversion have US GRAS status (137, 163).

Food categories and the authorized levels of use for steviol glycosides by regulatory authorities vary from one region to another. They generally include flavored and carbonated beverages, dairy products including fermented milk products, edible ices, table-top sweeteners, fruit and vegetable preparations, jams and jellies, cocoa and chocolate products, confectionary and chewing gum, a variety of sauces, breakfast cereals, some bakery products, processed fish products, foods for special dietary purposes, alcohol, several regional sweet and savory snack-based products, desserts, and food supplements (164, 165).

Stevia's primary advantage is that it is a plant-based sweetener of natural origin. There is no global definition or agreed-upon claim for the term “natural.” However, stevia leaf extract or steviol glycosides from the S. rebaudiana Bertoni plant are clearly defined as a natural sweetener in the food regulations of Korea, Malaysia, and Indonesia, and are reported as the “natural constituents” of the stevia plant in JECFA's 69th meeting report (26). The WHO, in its recent publication on reducing sugar in manufactured foods, also recognized stevia as a natural sweetener in its categorization of noncaloric sweeteners (i.e., natural and artificial) (166). It is generally acknowledged as a natural-origin sweetener in the United States and imagery and “natural” phraseology is used in many parts of the globe to convey to consumers the use of natural-origin plant-based stevia sweeteners. The labeling of steviol glycosides in the ingredient list of a food or beverage product can vary from one country to another. Examples include the following: stevia leaf extract, steviol glycosides, Reb A, rebiana, stevia, and in Europe, steviol glycosides (E960), etc.

Consumer insights and market trends

Across the globe, increased consumer awareness about the potential health benefits of reducing calories and sugar has resulted in a shift in consumer preferences for reduced-calorie/sugar foods and beverages, increasing the potential role of sugar substitutes in helping to address these preferences. In addition, an increasing interest in clean-label, organic, and natural LNCSs that do not compromise taste and function has helped to increase awareness about the benefits of stevia and increased demand for stevia-based products.

The global growth of stevia is estimated to exceed $1 billion by 2021 based on current market trends (167). The approval of high-purity stevia leaf extracts around the world has spawned hundreds of food and beverage launches. According to data accessed from Mintel's global products database, the number of products with stevia has grown considerably in the past 5 y (168). Since 2011 alone, ≥14,000 products were launched with stevia globally (Figure 3), and in 2016, 45% of the stevia-based products were in foods and 55% in beverages.

There is limited peer-reviewed research on consumer and health care professional perception and attitudes with regard to LNCSs. To determine aided-awareness, belief, and sentiment about LNCSs, including stevia, nationally representative population samples of ∼1000 adults, aged 18–64 y, from the United States, United Kingdom, Germany, China, India, Brazil, and Mexico were surveyed between 2011 and 2017 (169; PureCircle, proprietary data). Fifty percent of the respondents were men and 50% were women. The surveys contained ∼30 sweetener-related questions. The results indicated that across markets at initial launch, stevia awareness ranged from 8% to 35% and has increased to as high as 77% in Mexico (Figure 4). The increase in consumer awareness of stevia over time appears to correspond to the increases in product launches in a given country. In the same studies, participants were asked about their impression of stevia and their belief of stevia as a natural-origin, plant-based ingredient based on a 5-point Likert scale, which ranged from very positive to very negative (Figure 5). Positive responses (very positive to moderately positive) to the question on the overall impression of stevia ranged from 57% to 87% across several countries. The belief that stevia is “natural” product ranged from 48% to 86% across countries (Figure 5). There appeared to be a relation between overall impression of stevia and the belief that stevia is natural and vice versa (169).

An online beverage survey of 3361 US adults aged ≥18 y reported that <40% of participants identified added sugars as a primary concern when choosing beverages, despite dietary guidance to reduce added sugars in the diet (170). This study also reported a considerable level of consumer misunderstanding or confusion about the types of sugars in beverages. Another online study in the United Kingdom found that 65% of the participants reported no knowledge of the WHO sugar-intake guidelines (171). Subjects (77% female respondents) were asked to identify and classify 13 caloric sugars (added sugars) or LNCSs (aspartame and saccharin) on the food label, and only 4% correctly classified ≥10 from the ingredient lists. The authors noted that even well-educated consumers struggled to understand added sugars on food labels.

A study on the perception of LNCSs by dietitians from 5 European countries (France, Germany, Hungary, Portugal, and the United Kingdom) indicates that dietitians are uncertain, ambivalent, or have fears about adverse health effects of LNCSs (172). Their knowledge and opinion of LNCSs translated to varied approaches: some dietitians were undecided, some had the opinion that LNCSs should not be used, others felt that LNCSs should only be used as a transitional product, whereas another group recommended or at least allowed the use of LNCSs. Despite the lack of strong scientific evidence, some dietitians believed that sweet taste stimulates appetite. Uncertainty about possible adverse health effects or the safety of LNCSs, and distrust of the industry, were reasons why dietitians avoided recommending LNCSs. The authors of this study identified a clear need for authoritative positions and recommendations from appropriate and trusted sources as key to alleviating the ambiguity, uncertainty, and fear.

According to Euromonitor's July 2017 report on sugar and sweeteners, global consumers purchased 73 g total sugars/d in 2015, of which 22% was from table sugar, 19% from fruit (intrinsic sugar), and 16% from soft drinks (173). Sweet snacks, such as biscuits, snack bars, and confectionary, jointly provided >20 g sugar/d per capita in some of the high-sugar-consuming markets. Consumer perception is a critical factor, and according to Euromonitor, there appears to be a shift toward natural sweeteners, particularly natural, full-caloric sweeteners such as honey, coconut sugar, and brown rice sugar. According to Euromonitor, future development is expected to focus on natural sweeteners (173).

Authoritative positions on the use of nonnutritive sweeteners

Nutrition and health-related organizations such as the Academy of Nutrition and Dietetics, the American Heart Association (AHA), and the American Diabetes Association (ADA) currently have positions, scientific statements, or both that support the use of LNCSs, including stevia (74, 174). The Academy of Nutrition and Dietetics’ 2012 position paper graded the stevia data that they included in their evaluation as “fair” and the overall conclusion for LNCSs states that “consumers can safely enjoy a range of nutritive and nonnutritive sweeteners when consumed within an eating plan that is guided by current federal nutrition recommendations, such as the Dietary Guidelines for Americans and the Dietary Reference Intakes, as well as individual health goals and personal preference” (174). A 2012 joint scientific statement of the AHA and ADA on the use and health perspective of LNCSs, which included the review of evidence on stevia available at that time, concluded that, when used judiciously, LNCSs could facilitate reductions in added-sugar intake, thereby resulting in decreased energy intake and weight loss/weight control, with beneficial effects on related metabolic variables, as long as the substitution does not lead to consuming additional calories as compensation (74). In addition, the Council on School Health of the American Academy of Pediatrics in their position on snacks, sweetened beverages, and added sugar for schools also acknowledged the potential use of LNCSs for energy reduction in school-aged children (175). Furthermore, a recent expert panel in the United Kingdom concluded that natural-origin sweeteners, such as stevia, in blends with sugars offer consumers a way to help meet the UK recommendation of no more than 5% of energy from free sugars (176).

Although all major regulatory authorities around the world have approved and support the use of high-purity steviol glycosides in foods and beverages, policy positions and scientific statements on LNCS use similar to the ones by the Academy of Nutrition and Dietetics and the AHA and ADA are lacking in many other parts of the globe. This is a critical gap, because these statements offer actionable direction for practitioners and health care professionals who serve as an important and respected source of information and advice the public often needs. More research and education are needed to understand and help both consumers and health care professionals make informed choices based on credible scientific evidence.

Summary and conclusions

Several global and country-level authoritative dietary guidelines recommend a reduction in added-sugar intake due to the growing prevalence of overweight, obesity, and diabetes around the world. These guidelines include recommendations to keep added-sugar intake to <10% of total calorie intake, and as low as 5% for additional health benefits according to the WHO (59) and the UK Scientific Advisory Commission on Nutrition (60). The replacement of caloric sweeteners in foods and beverages with high-purity stevia leaf extract sweeteners (i.e., steviol glycosides) is a useful and cost-effective tool in reducing added-sugar intake.

Natural-origin steviol glycosides are the natural, sweet constituents of the leaves of the S. rebaudiana Bertoni plant, which remain unaltered during extraction and purification. The safety of consumption of high-purity steviol glycosides at or below the ADI is well established. Although there are opportunities for additional research as outlined in sections of these proceedings, evidence to date shows that steviol glycosides are safe, noncariogenic, and nonhypertensive and have minimal impact on the gut microbiota. Human studies have reported no negative gastrointestinal side effects. When used to displace carbohydrate and sugar in the diet, studies with high-purity steviol glycosides in healthy individuals and those with diabetes support a reduction in postprandial blood glucose as well as reduced sugar and energy intake. There is no evidence that shows an increase in appetite for sugar or sweet products when LNCSs or stevia-containing foods are consumed. Therefore, stevia leaf extract sweeteners are a beneficial and critical tool in sugar and calorie reduction, diabetes, weight management, and healthy lifestyles. Recent innovations have resulted in better-tasting, natural-origin, high-purity stevia leaf extracts that help both product developers and consumers make the switch from full-calorie/sugar products to reduced or zero-calorie/sugar-added products to assist in meeting dietary guidelines consistent with current health and nutrition policy recommendations.

Acknowledgments

We thank Khor Geok Lin and Margaret Ashwell, members of the PureCircle Stevia Institute (formerly GSI) advisory board for their contributions to the Stevia Symposium. We also thank Ashi Okonneh who helped with the Mintel data and PureCircle consumer survey figures and John Martin for support on PureCircle's sensory data. The authors' responsibilities were as follows—PS: conceptualized and led the stevia symposium and proceedings and developed the figures; PS and KTA: chaired and co-chaired the symposium, respectively; KTA, BAM, UW-R, PBJ, PJR, IR, and PS: gave presentations at the symposium; PS and RM: edited the manuscript and developed the tables; and all authors: contributed to writing the manuscript and read and approved the final manuscript.

Notes

Published in a supplement to The Journal of Nutrition. Presented at the symposium “Stevia Leaf to Stevia Sweetener: Exploring Its Science, Benefits, and Future Potential,” held in Chicago, Illinois, 22 April 2017, at the ASN 2017 Experimental Biology conference. The symposium was organized by the Global Stevia Institute (GSI) and funded by PureCircle, Inc. The contents are the sole responsibility of the authors. The article comprising this supplement was developed independently to provide a comprehensive review of stevia. The Supplement Coordinator for this supplement was Priscilla Samuel (PureCircle Stevia Institute; supported by PureCircle Ltd.). Supplement Coordinator disclosure: Dr. Samuel serves as the head of the Global Stevia Institute, now known as the PureCircle Stevia Institute, and is an employee of PureCircle Ltd., and the Global Stevia Institute is supported by PureCircle Ltd., a leading supplier of stevia ingredients. Publication costs for this supplement were defrayed in part by the payment of page charges. This publication must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact. The opinions expressed in this publication are those of the authors and are not attributable to the sponsors or the publisher, Editor, or Editorial Board of The Journal of Nutrition.

Author disclosures: PS heads the Global Stevia Institute (GSI) now known as the PureCircle Stevia Institute and is employed by PureCircle, Inc. All of the speakers (KTA, BAM, UW-R, PBJ, PJR, and IR) received travel expenses and an honorarium from GSI for their participation in the Stevia Symposium held at the ASN 2017 Experimental Biology conference, April 2017. RM received travel expenses for attending the conference and fees for assisting with editing the manuscript from GSI. KTA, is a GSI advisory board member and is a consultant to the Calorie Control Council. At the time of the symposium, BM was a GSI advisory board member and currently is a consultant to the Calorie Control Council. UW-R is an advisory board member of GSI and the European Stevia Association (EUSTAS) and has received research funding from both GSI and EUSTAS. PBJ is an honorary member of EUSTAS and a full voting member since 2009. PJR has received research funding from Sugar Nutrition UK and consultant fees from Coca-Cola Great Britain and the International Sweeteners Association. IR has received speaker fees from the Calorie Control Council.

Publication costs for this supplement have been provided by the PureCircle Stevia Institute (formerly GSI), which is funded by PureCircle Inc.

Supplemental Tables 1 and 2 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.

Abbreviations used:

 
• ADA

  American Diabetes Association

 
• ADI

  Acceptable Daily Intake

 
• AHA

  American Heart Association

 
• EDI

  Estimated Daily Intake

 
• EFSA

  European Food Safety Authority

 
• ESL

  first water extract

 
• FSANZ

  Food Standards Australia New Zealand

 
• GRAS

  Generally Recognized As Safe

 
• HbA1c

  glycated hemoglobin

 
• HFCS

  high-fructose corn syrup

 
• JECFA

  FAO/WHO Joint Expert Committee on Food Additives

 
• LNCS

  low- and no-calorie sweetener

 
• NOAEL

  No Observed Adverse Effect Level

 
• Reb

  rebaudioside

 
• SE

  steviol equivalent

 
• SL

  dried stevia leaves

 
• SLE95

  stevia leaf extract with ≥95% purity

 
• TRPM5

  transient receptor potential cation channel subfamily M member 5

References