点击显示 收起
1 From the Department of Anthropology, University of Auckland, Auckland, New Zealand.
2 Address reprint requests to B Turnbull, Sensory Science Research Centre, University of Otago, PO Box 54, Dunedin, New Zealand. E-mail: bianca.turnbull{at}stonebow.otago.ac.nz.
ABSTRACT
Background: Understanding what motivates the preference for and selection of foods has important health implications. Research suggests that the phytochemicals present in green leafy vegetables contain anticarcinogenic properties. As a result of the bitter taste of phytochemical compounds, however, foods containing these are often not well accepted, particularly by children.
Objective: We aimed to study the relation between sensitivity to the bitter taste of 6-n-propylthiocuracil (PROP) and acceptance of bitter- and strong-tasting foods in 3–6-y-old children.
Design: Two independent procedures, a threshold detection and a suprathreshold intensity task, were used to measure individual sensitivity to PROP, and 3 independent tasks were used to assess food acceptance.
Results: Sensitivity to the bitter taste of PROP was positively correlated with dislike of the taste of raw spinach (P < 0.05).
Conclusions: The acceptance of spinach may to some extent be mediated by individual taste perception and be predictable via both threshold and suprathreshold measures of PROP taste sensitivity. Furthermore, children as young as 3 y of age can partake in direct investigations of taste, reliably comply with test procedures, and accurately communicate taste perceptions and preferences under study conditions.
Key Words: 6-n-Propylthiouracil • taste sensitivity • food acceptance • children • phytochemicals • spinach
INTRODUCTION
The mouth is a route through which nutrients and toxins alike enter the body. It is effectively the final point where one decides to accept or reject food items (1). As omnivores, what we eat is closely related to survival; therefore, the ability to distinguish edible from inedible, avoid toxins, and eat a balanced diet must be established early (2). Taste mechanisms appear to be functionally mature early in gestation (3), and human infants can distinguish sweet and bitter stimuli at birth (4–7). As a child matures and gains some degree of autonomy over food choice, multiple factors affect what is preferred and ultimately eaten. Although taste remains an important determinant of food acceptance during childhood, it also appears to be the primary dimension of food rejection at this time (1, 3). Rejection of bitter- and strong-tasting foods is an especially salient feature of childhood (8), at which time bitter-tasting green vegetables often appear at the top of a child’s most disliked food list (9, 10).
Sensitivity to the bitter taste of phenylthiocarbamide (PTC) and the related thiourea compound 6-n-propylthiouracil (PROP) is a heritable trait. Since discovering the hereditary nature of taste sensitivity in 1931 (11), researchers have used these synthetic chemicals to both classify and comprehend the complexities of taste sensitivity. Both PTC and PROP have a common thiocyanate moiety, N-C=S (12). Glucosinolate compounds that occur naturally in cruciferous vegetables (broccoli, Brussels sprouts, cabbage, and kale) share this chemical grouping (13) and, additionally, a common taste quality: bitterness.
Although it has long been known that sensitivity to PTC and PROP is related to enhanced perception of other bitter compounds (14–17), sensitivity to PROP also appears to be related to the acceptance of bitter-tasting compounds and the foods in which they appear (18–21). The association between PTC-PROP taste sensitivity and the acceptance of cruciferous vegetables appears particularly strong. Those with lower PTC taste thresholds appear to have more food dislikes than do less-sensitive individuals (20, 21) and prefer a greater number of mild- over strong-tasting foods (21).
Although these results bear on the salience of taste mechanisms in the perception and acceptance of bitter- and strong-tasting foods among adults, few have investigated the relation between taste sensitivity and food acceptance in children. Anliker et al (22) were the first to assess the relation between PROP taste sensitivity and food acceptance in children aged < 7 y. In their study, PROP nontaster children (aged 5–7 y) selected Cheddar cheese significantly earlier, and whole milk significantly later, than did PROP taster children.
At the time of this research, no study had looked at the relation between PROP taste sensitivity and food acceptance in children aged < 5 y. Although our study did include children aged 6 y, most of the participants were aged 4 y at the time of testing. By directly assessing taste sensitivity in children predominantly aged < 5 y, we hoped to attain an important insight into the mechanisms responsible for early food acceptance and rejection.
SUBJECTS AND METHODS
Subjects
The participants in this study were 42 children, 22 boys and 20 girls, aged 3–6 y (
Data collection
Data were collected over 2 testing sessions, each lasting between 20 and 30 min. The researcher and participant were alone during these testing sessions, which took place on weekday mornings in an area of the school both familiar to the child and practical for research purposes. Before testing, each child was instructed to demonstrate the ability to follow a sip-and-spit procedure by spitting distilled water into a bowl. The testing protocol adopted in this study closely followed that outlined by Anliker et al (22). Some procedures in this study were modified to accommodate the younger participants. All procedures were approved by and conducted in accordance with the University of Auckland Human Subjects Ethics Committee.
Measures of taste sensitivity
Individual taste thresholds were established by using a forced choice–detection threshold procedure (23). Fifteen serial dilutions of PROP were prepared from a master solution and ranged from 0.0051 mmol/L distilled water (solution 1) to 3.2 mmol/L distilled water (solution 15). The children tasted PROP solutions in ascending order, relative to distilled water. All solutions were presented in identical 35-mL medicine cups, and participants were instructed to rinse with water before and after tasting the test solutions.
To begin, the children were invited to taste 10 mL of the most dilute PROP solution alongside a sample of distilled water. These solutions were presented in balanced order and the children were specifically asked to point to the solution believed to taste like water. Solution concentration increased until participants were able to correctly distinguish the 2 solutions.
Actual testing began with the presentation of the next weakest PROP solution, paired again with distilled water. If an incorrect choice was made at this concentration, testing continued as before with the next strongest solution. Alternatively, a correct choice led to the presentation of the next weakest PROP solution. Testing continued until 4 reversals had been completed, ie, either an incorrect choice followed 2 correct choices or a correct choice followed 2 incorrect ones. Individual PROP thresholds were calculated by taking the mean of the last 4 reversals (24).
To assess suprathreshold taste, the children were asked to taste and rate the intensity of 10 solutions (10 mL). Four of these solutions were suprathreshold solutions of PROP (0.056, 0.18, 0.56, and 1.8 mmol/L), 4 were aqueous solutions of sodium chloride (0.032, 0.1, 0.32, and 1.0 mmol/L), and the remaining 2 were distilled water. The 4 salt solutions were used as standard stimuli. Although it is acknowledged that not all persons evaluate the taste of sodium chloride identically, it was expected that subject response would be more similar for the standard relative to the test stimuli (25).
A simple ball and bucket game was used to record participant intensity ratings. Three buckets were placed at different heights to exemplify increasing taste intensities. The scale was explained to the children, who were instructed to place a table tennis ball into the appropriate bucket to communicate the perceived intensity of each solution. The bucket at the bottom denoted the mildest intensity, that which could be equated to water. The bucket above this represented solutions that tasted "quite strong," and the highest bucket, "very strong." After they tasted each solution, the children were instructed to rate the intensity of the solution by placing a table tennis ball into the appropriate bucket. Each child was also asked to verbally express their perceptions to the researcher to ensure correct use of the scale.
Measures of food acceptance
At the second testing session, individual food acceptance was assessed by using 3 independent tasks. The order-by-choice procedure (26) called for 7 food and beverage items to be tasted and ranked from most to least preferred. These items were raw spinach, raw broccoli, cooked broccoli, banana, lemonade (0.3 mol glucose/L), whole milk, and Cheddar cheese. This procedure is effective because it forces individuals to rank taste preferences with reference to other items presented simultaneously. It has proven a reliable measure of food acceptance in both preschool (26) and school-aged children (22).
The 7 foods were presented in whole form to assist recognition. The children were first asked to name each item and then taste a portion of each. Once each item had been tasted and named the children were asked to point to the item that tasted the best. This item was subsequently removed and the procedure repeated until all 7 items had been given an order of choice.
The second measure of food acceptance required the children to taste and hedonically rate the same 7 foods presented previously on the order-by-choice task. Hedonic ratings were recorded immediately after each item was tasted by using 3 small buckets. These 3 buckets were painted with either a smiling, neutral, or frowning face on the side, representing good taste ("yummy"), "OK taste," and bad taste ("yucky"), respectively. Immediately after each item had been tasted, the children were asked to place a table tennis ball into the appropriate bucket to convey their hedonic ratings. Verbal descriptions of the taste were also encouraged to ensure correct understanding and use of the hedonic scale.
The same hedonic scale was used for the third measure of food acceptance, the verbally administered questionnaire. For 30 food and beverages, the children were asked if they had tasted a specific item previously and, if so, whether they thought the item tasted "yummy," "yucky," or "just OK." The foods were predominantly sweet and bitter items. Particular attributes and preparation styles were conveyed to the children (ie, cooked cabbage, chocolate milkshake). Five foods previously tasted on both the order-by-choice and hedonic rating procedure were again presented on the verbally administered questionnaire. These items were raw broccoli, raw spinach, lemonade, whole milk, and Cheddar cheese.
Analysis
In an effort to provide a more robust measurement of taste sensitivity, mean PROP taste thresholds and PROP intensity ratings were not used to place participants into the conventional nontaster, taster, and supertaster groups. Although this clearly restricted our ability to compare our results with previous research efforts (22), it was hoped that a continuous distribution of scores would enable better manipulation of the data and give greater strength to the results. All data were analyzed by using SPSS 8.0 for WINDOWS (SPSS Inc, Chicago). Mean PROP taste threshold scores and food acceptance, as measured by the order-by-choice, hedonic rating, and verbally administered questionnaire tasks, were evaluated by linear regression analysis.
Of the 4 PROP solutions administered on the suprathreshold intensity rating task, ratings for only 1 (1.8 mmol/L), were used for analysis. PROP intensity ratings and ratings on the 3 food acceptance tasks were analyzed by Pearson chi-square statistic. The consistency of responses made across the 3 measures of food acceptance was evaluated by a two-tailed Spearman’s correlation.
RESULTS
PROP taste thresholds and PROP intensity ratings remained independent of each other during analysis and were individually assessed relative to the 3 measures of food acceptance. Individual mean PROP threshold scores and intensity ratings for the single PROP suprathreshold solution are presented in Figure 1 and Figure 2, respectively.
FIGURE 1. . Distribution of 6-n-propylthiouracil (PROP) detection thresholds ranging from 0.0051 mmol/L distilled water (no. 1) to 3.2 mmol/L distilled water (no. 15).
FIGURE 2. . 6-n-Propylthiouracil (PROP) taste intensity ratings for the suprathreshold solution containing 1.8 mmol PROP/L distilled water. PROP taste intensity ratings were as follows: tastes like water (no. 1), tastes "quite strong" (no. 2), and tastes "very strong" (no. 3).
A linear regression analysis of individual PROP threshold scores relative to the rank ordering of preferences did not show PROP threshold scores to be a predictor of food acceptance. Furthermore, rank ordering of foods and beverages on the order-by-choice task was not related to PROP intensity ratings.
Independent analysis of PROP taste thresholds and suprathreshold intensity ratings relative to hedonic ratings for the same 7 foods showed that those individuals sensitive to the bitter taste of PROP were more likely to dislike the taste of raw spinach (P < 0.05). The results of this analysis are presented in Table 1.
View this table:
TABLE 1 . P values for the analysis of independent measures of 6-n-propylthiouracil (PROP) taste sensitivity and the hedonic ratings for 7 foods and beverages
Of the 30 items on the verbally administered questionnaire, ratings for only 1 item, raw spinach, were significantly correlated with PROP taste sensitivity (P < 0.05). The questionnaire results for 5 items previously presented in the order-by-choice and hedonic rating tasks are shown in Table 2. A two-tailed Spearman’s correlation showed ratings for these 5 items to be consistent across all 3 measures.
View this table:
TABLE 2 . P values for the analysis of independent measures of 6-n-propylthiouracil (PROP) taste sensitivity and the hedonic ratings for 5 items on the verbally administered questionnaire1
DISCUSSION
In accordance with the results of previous research efforts (26), the results of this study suggest that children under the age of 5 y are able participants in taste research and can accurately communicate taste perceptions and preferences under study conditions. It is recognized, however, that the accuracy of the results attained from this age group is largely dependent on the ability of the researcher and the procedure adopted to account for the limitations of younger research participants.
Children are renowned for their rejection of bitter-tasting green vegetables. However, the results of this study did not reflect group homogeneity with regard to vegetable acceptance. Instead, our results showed that the acceptance of some vegetables containing bitter-tasting compounds may be mediated, and to some extent predicted, by measures of PROP taste sensitivity. Of particular interest is the discovery that PROP taste sensitivity may predict the acceptance of vegetables not exclusive to the Cruciferae family. A study (BJ Tepper, L Steinmann, unpublished observations, 1998) reported that taster girls (aged 4–5 y) disliked the taste of raw broccoli significantly more than did nontaster girls. Although broccoli and other cruciferous vegetables contain glucosinolate compounds with a chemical grouping (N-C=S) similar to that in both PTC and PROP, spinach does not belong to the same plant family and, consequently, does not share this thiocyanate moiety. Spinach is a member of the Chenopodiaceae family and the bitter-tasting leaves of this plant are particularly high in carotenoids, such as lutein and ß-carotene, which are proven antioxidant agents (27). In accordance with our findings, Tepper and Steinmann (unpublished observations) found no relation between PROP taste sensitivity and the preference for cooked broccoli. This may be due to the breakdown of glucosinolate compounds in broccoli during cooking. When the temperature of cruciferous vegetables is increased through heating, their chemical makeup and volatile flavor may be significantly altered (28).
Regarding the methods adopted to study food acceptance, it has been suggested that verbally administered questionnaires are unreliable and instead reflect individual attitudes toward the verbal concept of the foods (29). Despite having a considerable number of bitter-tasting fruit and vegetables present on the verbally administered questionnaire, only the preference ratings for spinach showed any correlation with PROP taste sensitivity. This suggests that young children may be better able to communicate their taste preferences through direct testing procedures. In addition, and certainly in the case of younger subjects, reported taste dislikes on verbally administered questionnaires may in fact reflect dietary unfamiliarity and not actual taste preferences (30). Certainly, the results of this study attest to the salience of actual tasting when investigating food acceptances of young children.
Awareness of what mediates the acceptance of foods, particularly green leafy vegetables, at an early age has important health implications. It has long been known that the compounds present in green leafy vegetables are active defenders against carcinogenic cell growth, and, like broccoli, spinach contains important antioxidant compounds capable of neutralizing cancer-causing free radicals (31). The inclination for children to consume a diet rich in energy and yet remain nutritionally deprived is a pattern reflected in many health surveys (32). Many nutritional messages tend to focus largely on promoting the nutritional qualities of foods with the premise that by improving health knowledge, healthy eating habits will follow (33). The implicit assumption that healthy eating knowledge translates into actual behavior, however, is inherently flawed (34). Multiple factors influence the development of individual eating patterns, and, most importantly, children, like adults, have unique genetic, social, and cultural histories that dictate their food choice decisions. With increased research attention, taste sensitivity data may provide an important marker for the early identification of children at risk of developing adverse eating patterns and, subsequently, diet-related diseases.
ACKNOWLEDGMENTS
We extend our thanks to John Prescott for his advice and guidance.
REFERENCES