Amid growing concerns over excessive sugar intake, steviol glycosides, especially rebaudioside A (Reb A) and rebaudioside M (Reb M), have emerged as promising natural alternative sweeteners. To replace sucrose effectively, it is critical to estimate t...
Amid growing concerns over excessive sugar intake, steviol glycosides, especially rebaudioside A (Reb A) and rebaudioside M (Reb M), have emerged as promising natural alternative sweeteners. To replace sucrose effectively, it is critical to estimate the relative sweetness of these compounds at sucrose-equivalent concentration (SEC). However, the inherent bitterness and off-flavors of Reb A and Reb M can suppress sweetness perception, complicating accurate estimation. Additionally, individual differences in sensory perception, influenced by genetic factors such as sensitivity to 6-n-propylthiouracil (PROP), and behavioral factors such as sweet liking and stevia consumption frequency, introduce inter-individual variation, thereby influencing the determination of SEC and consumers’ sweetness perception. This study aimed to empirically examine, among general consumers, how PROP sensitivity and dietary behavior traits affect the relative sweetness (RS) and sensory profiles of Reb A and Reb M.
This study involved 143 adults aged 20-60 years over a three-week period. RS and sensory profiles of Reb A and Reb M were determined at 4% and 8% SEC based on the Korean criteria for low-sugar labeling and the typical concentrations used in commercially available beverages. A sucrose-caffeine mixture was used as a control stimulus to clearly assess the influence of bitterness. Tests were conducted three times, one test per week. In the first and second weeks, RS was estimated using a two-alternative forced choice (2-AFC) test, with sample order randomized via a Williams Latin square design. In the third week, sensory profiles of the samples at 4% and 8% SEC were obtained using a 7-point scaled rate-all-that-apply (RATA; 1 = very weak, 4 = neither weak nor strong, 7 = very strong). Participants’ PROP sensitivity was determined by rating a 3.2 mM PROP solution twice using the generalized labeled magnitude scale (gLMS). In addition, demographic and behavioral data such as sweet liking and stevia consumption frequency were collected. Participants were classified according to the quartile distribution of PROP intensity ratings, and the preliminary classification was further refined using kernel density estimation, which eliminated 37 individuals located near the quartile cutoffs, leaving a total of 106 participants. Sweet liking was grouped using predefined ranges on the 7-point hedonic scale (LSL: 1-3; MSL: 4-5; HSL: 6-7), and stevia consumption frequency was classified as low (rarely/never or 1-2 times/6 months) versus higher frequencies based on self-reported intake. RS values were estimated as the concentration corresponding to a 0.5 response proportion on the concentration-response curve. However, for the sucrose-caffeine control, one-sided binomial test identified concentrations at which sweetness differed significantly from sucrose reference. Group differences in RS were examined using a binomial generalized linear model (GLM). Sensory profiles were analyzed through multivariate analysis of variance (MANOVA), principal component analysis, and multiple factor analysis to comprehensively characterize complex interactions among factors. Significant MANOVA attributes were further analyzed using analysis of variance with Duncan’s test or paired/independent-samples t-tests.
Results showed that participants were classified, according to PROP sensitivity, as non-tasters (n = 30), medium tasters (n = 46), and supertasters (n = 30). Based on sweet liking, they were grouped into low (n = 28), medium (n = 52), and high sweet likers (n = 26), while stevia consumption frequency resulted in two groups (low stevia consumption frequency group, LSC; n = 46 and high stevia consumption frequency group, HSC; n = 60). Contrary to expectations, PROP sensitivity had no significant effect on the RS and RATA results for Reb A and Reb M. In contrast, sweet liking and stevia consumption frequency had significant impacts on both RS and RATA results. At 4% SEC, low sweet likers showed higher RS values and higher RATA ratings for the sweet aftertaste of both Reb A and Reb M than other groups, whereas medium and high sweet likers perceived non-sweet attributes, such as chalky mouthfeel, more intensely. This suggests that higher sweet liking, accompanied by greater habitual exposure to sweetness, may induce sensory adaptation, reducing sweetness dominance and revealing masked non-sweet characteristics. Stevia consumption frequency produced clear differences in both 2-AFC and RATA results. The LSC showed a significantly higher baseline on the C-R curves than the HSC for Reb M and Reb A at both 4% and 8% SEC. Additionally, in the RATA data the HSC provided relatively higher ratings for non-sweet attributes such as bitter aftertaste and off-flavors. These effects varied with the concentration-dependent changes in the sweeteners’ sensory characteristics: For Reb M at 4% SEC, group differences were significant, but at 8% SEC, strong sweetness masked bitterness, eliminating differences between groups. For Reb A, whose bitterness increased with concentration, the HSC clearly distinguished bitter aftertaste from sweetness at 8% SEC, resulting in pronounced between-group differences in perceived sweetness. These differences may be attributable to adaptation to sweetness and perceptual learning, whereby the HSC with greater product experience more clearly distinguished sweetness from other sensory attributes, whereas the LSC tended to conflate sweetness with non-sweet sensations, inflating sweetness ratings.
In conclusion, this study demonstrates that consumers’ sweetness perception and evaluation of Reb A and Reb M are driven more by behavioral factors than by genetic factors such as PROP sensitivity. By demonstrating that dietary behavior factors, rather than PROP sensitivity alone, influence two key indices of Reb A and Reb M sweetness, relative sweetness and sensory profiles, this study helps clarify major sources of variability in sweetness evaluation. These findings provide a solid basis for future research aimed at understanding individual differences in sweetness perception and offer practical implications for the application of rebaudiosides in product development.