Genetic taste sensitivity to 6-n-propylthiouracil influences food preference and reported intake in preschool children

Appetite (2002) 38, 3±12 doi:10.1006/appe.2001.0441, available online at http://www.idealibrary.com on

1

Original Article

Genetic taste sensitivity to 6-n-propylthiouracil influences food preference and reported intake in preschool children Kathleen L. Keller, Lone Steinmann, Ricky J. Nurse and Beverly J.Tepper Department of Food Science,Cook College, Rutgers University, New Brunswick, NJ,U.S.A. (Received12 January 2001, finalrevision 26 October 2001, accepted in revised form 4 November 2001)

Adult tasters of 6-n-propylthiouracil (PROP) are more sensitive to bitter taste and fattiness in foods, and often show lower acceptance of foods that are high in these taste qualities. This study hypothesized that PROP taster children would show lower acceptance of these same foods. Sixty-seven preschool children were classified as PROP tasters (N ˆ 43) or nontasters (N ˆ 24) using a suprathreshold screening solution. Children rated acceptance of 10 bitter and/or fat-containing foods using a 5-pt. facial scale. Parents completed a food frequency questionnaire to estimate their child's intake. Taster children showed lower acceptance of raw broccoli and American cheese ( p  0
05). Taster-girls showed lower acceptance of full-fat milk than nontaster-girls ( p  0
05). This effect was not seen in boys. Nontasters reported more daily intake of discretionary fats than tasters ( p  0
05), an effect largely due to nontastergirls, in whom reported intake was 2±3 more servings per day than taster-girls, and boys of both groups. These data suggest that PROP taste sensitivity plays a role in acceptance of certain bitter cruciferous vegetables and cheese by young children. In addition, taster group differences in acceptance of full-fat milk and intake of discretionary fats seen in girls, suggest that gender-specific environmental factors might interact with genetics to influence fat preferences. # 2002 Elsevier Science Ltd

Introduction Children in the U.S. consume too many fats and sweets and not enough fruits and vegetables, dietary practices that have been associated with a growing epidemic of childhood obesity and an increased risk of other chronic diseases later in life (Munoz et al., 1997; Krebs-Smith et al., 1996). In order to improve these dietary habits, it is first necessary to understand the factors that contribute to the development of children's food preferences. Certain factors that influence food

Address correspondence to: Dr Beverly J. Tepper, Department of Food Science, Cook College, Rutgers University, 65 Dudley Road, New Brunswick, NJ 089018520, U.S.A. E-mail: [email protected] 0195±6663/02/010003+10 $35.00/0

preferences are genetically predisposed. For example, at birth, humans prefer sweet (Cowart, 1981) and reject bitter tastes (Rozin & Vollmecke, 1986). Additionally, food neophobia, or the ``fear of new foods'' is common to humans, and might be protective against the ingestion of toxic or poisonous substances (Rozin & Vollmecke, 1986). Furthermore, the ability to learn food preferences based on their association with various environmental cues is also predisposed. To highlight this, Birch and colleagues have shown evidence for conditioned food preferences in children (Birch et al., 1990). In this study, children learned to prefer flavours paired with high-calorie drinks, but had lower preferences for flavours paired with low-calorie drinks. These genetic predispositions are integral factors in shaping the development of food preferences in young children. # 2002 Elsevier Science Ltd

4

K. L. Keller et al.

Despite their genetic similarities in basic taste responses, it is well known that humans vary greatly in their preferences for individual foods. Certainly, even within the same family, one child may readily accept bitter vegetables such as broccoli or spinach, whereas a sibling may reject the same foods. In studying how children's food preferences are formed, environmental factors must be considered. Included in these factors are the familiarity a child has with a food (Birch et al., 1987), the social context in which the food is served (Birch et al., 1980), and both peer and parental interactions (Burt & Hertzler, 1978; Birch, 1979). Mere exposure to foods through ingestion has been shown to be an effective practice to increase food liking (Birch et al., 1987). Birch et al. have carried out extensive work to suggest that parents who exhibit high levels of control over their child's food intake might increase their child's preferences for foods that are restricted in the home environment (Fisher & Birch, 1999). Finally, using foods as rewards has been shown to increase a child's preference for that food (Birch et al., 1980). This work has resulted in a theory by which children learn to prefer palatable high-fat and sugary foods because they are so commonly used by parents and teachers as rewards. This ability to learn preferences for foods based on a particular environmental situation is alarming, given that palatable, energy-dense foods are so abundant and readily available in developed countries. Recent findings on the genetics of taste may provide additional insight into explaining individual differences in food preferences. The ability to taste bitter thiourea compounds, including 6-n-propylthiouracil (PROP) and phenylthiocarbamide (PTC) is inherited. Approximately 70% of the U.S. Caucasian population are sensitive to PROP at low concentrations (tasters), while the remaining 30% are taste blind to PROP (nontasters) (Blakeslee, 1931; Snyder, 1931; for review see Tepper, 1998). PROP tasters are more sensitive to many oral sensations, including bitter and sweet tastes and the sensation of fats (Bartoshuk, 1979; Bartoshuk et al., 1988; Looy & Weingarten, 1992; Tepper & Nurse, 1997). PROP tasters generally dislike or show ambivalence to foods with high concentrations of these taste qualities, whereas PROP nontasters tend to prefer them (Glanville & Kaplan, 1965; Duffy & Bartoshuk, 2000). For example, some studies in adults have shown that PROP tasters have lower preferences for bitter citrus fruits such as grapefruit (Drewnowski et al., 1997) and cruciferous vegetables (Drewnowski et al., 2000). In addition, Tepper and colleagues recently showed that PROP tasters were better able to discriminate between high-fat (40% fat) and low-fat (10% fat) salad dressings, although they showed no

preference for either. Nontasters, in contrast, preferred the high-fat sample (Tepper & Nurse, 1997, 1998). Data examining the relationship between PROP status and food preferences in children are sparse. Anliker et al. (1991) studied the food preferences of 5±7-year old children and found an association between taster status and liking of cheese and milk. Nontasters preferred cheese to other foods, whereas tasters preferred milk. Anliker et al. (1991) interpreted their findings based on the perceived bitterness of these foods. However, in light of recent studies suggesting that taster status might influence fat perception in foods, further investigation of how taster status influences a child's acceptance of both bitter and fat containing foods is warranted. Moreover, previous work in children did not examine dietary intakes of tasters and nontasters. Establishing the relationship between PROP sensitivity and dietary intake is critical for determining the nutritional significance of the PROP phenotype. The objectives of the present study were twofold. The first objective was to classify children as tasters or nontasters of PROP and measure their acceptance of foods that are both familiar to the children and vary in bitter taste and fat texture properties. Bitter foods were tested to confirm earlier findings in children (Anliker et al., 1991); fat containing foods were tested to investigate the association between taster status and fat preference, previously observed in adults with salad dressings (Tepper & Nurse, 1998). The second objective was to determine the relationship between PROP taster status and the children's reported intake of bitter and fat-containing foods. This study tested the hypothesis that taster children would have both a lower acceptance and reported intake of foods with stronger bitter taste and increased fat content.

Materials and methods Subjects Sixty-seven, 4±5-year-old children (33 girls and 34 boys) from the Rutgers University Nutritional Sciences preschool participated in this study. Data were collected over a 3-year period, and approximately 25 children were tested each year. This study was approved by the Institutional Review Board at Rutgers University and written informed consent was obtained from parents to allow their children to participate in the study.

PROP taster status PROP taster status was determined by using a method developed by Lawless (1980). In this method, subjects

Genetic taste influences in children

sip a single solution of 0
56 mM PROP and rate the intensity of the solution on an eight point scale ranging from 0 (no taste) to 7 (very strong taste). We used the same PROP concentration as that used by Lawless (1980), but simplified the sample presentation to a two-alternative response where children responded ``yes'' or ``no'' when asked if they could taste the solution. More specifically, the children tasted 10 mL of a 0
56 mM PROP (6-propyl-2-thiouracil, Aldrich Chemical Inc., Milwaukee, WI) solution in spring water, after which they were verbally asked the question ``Do you taste anything?''. Children who responded ``no'' were then asked if the solution tasted like water. If they reported that the solution had no taste and tasted like water, they were recorded as nontasters. Those who reported a taste for the solution were further questioned as to what the solution tasted like. Responses of ``bad,'' ``bitter,'' ``sour,'' and ``yucky'' were all recorded as tasters. If a child did not verbally state that the solution had a taste, but exhibited classic rejection signs, such as grimmacing or frowning, this child was recorded as a taster. Two children gave ambiguous responses. In both cases, children reported that the solution had a taste, but that it tasted ``good.'' Both children were retested at a later date, at which time they reported that the solution tasted ``good,'' but like water. These children were classified as nontasters.

Taste stimuli Children often reject foods that are unfamiliar (Birch & Marlin, 1982). To avoid this bias, we chose common foods that the children had previously been offered as part of daily snacks at the preschool. Foods with predominant taste characteristics of either bitter or fat were tested. Foods that varied in bitterness were: raw and cooked broccoli flowerets, orange juice (Tropicana Products Inc., Brandenton, FL) and orange-grapefruit juice (a 2 : 1 mixture of orange juice and Ruby Red Grapefruit Juice, Tropicana Products, Inc.), semisweet and milk chocolate morsels (Hershey Foods, Hershey, PA.), and American cheeses (Deli Deluxe American slices and Old English Sharp American slices, Kraft Inc., Northfield, IL). To test for fat acceptance in foods, we used: full-fat and fat-free milk (Grand Union Inc., Wayne, NJ), and beef (high-fat) and turkey (low-fat) hotdogs (Ballpark Inc. Detroit, MI). Approximately 10 g of each sample was placed into small souffle cups for tasting. Broccoli was cooked by covering 100 g of the sample with plastic wrap and micowaving on high power for 2 minutes. Hot dogs were boiled in spring water for 1 minute and served to the children at room temperature.

5

Procedure Testing was conducted at the Nutritional Sciences Preschool. Each child was tested separately by a researcher in a 15±20 minute one-on-one session. Children who were in the 4-year-old classes were tested in a separate kitchen, adjacent to the preschool classroom. The kitchen was equipped with child size tables and chairs for the testing procedure, and children were familiar with the room, as it was also the place where they were served daily snacks. We were prohibited from removing the 3-year-olds from the class, so these children were tested in a corner of the classroom, isolated from other children and teachers. Before laboratory taste tests were begun, the researchers visited the preschool classes during playtime to establish familiarity with each child. On the day of the test, the researcher explained to the child that they were ``Going to play a food game.'' The child was instructed on how to rate food preferences using a 5-point facial hedonic scale that has previously been validated for children of this age (Chen & Resurreccion, 1996). The scale uses 5 faces with expressions that range from smiling, to neutral, to frowning. To ensure that each child understood the scale, the researcher asked the child to name his/her favorite food. The researcher then asked the child ``Which smiley face would you point to if I gave you some of (child's favorite food)?'' The same process was repeated for the child's least favorite food. The procedure was explained again if the child still did not understand how to use the scale. After two explanations, if the child still seemed confused, he or she was allowed to choose a toy and further testing was not done. Out of a potential 73 children, 5 did not understand the scale and were not used in the study. Foods for the taste test were prepared 1-hr before testing, and presented to the children one at a time, in random order. The child used the scale to show how much he or she liked the test food, and the researcher recorded the child's responses. Children were instructed to rinse their mouths with water between tasting each food. To keep children from losing interest, a poster board was used for each child with pictures of all the test foods. After tasting each food, the child was allowed to place a sticker on the picture corresponding to that food, according to a procedure adapted from Kimmel et al. (1994). PROP tasting was conducted at the end of the session.

Food frequency A food frequency questionnaire was used to estimate habitual intake of the children. The questionnaire estimates the number of standard servings consumed

6

K. L. Keller et al.

daily or weekly for 141 common foods or food groups. The questionnaire had been validated in adults against 4-day diet records for energy, macronutrient intake, and frequency of consumption (Tepper, unpublished 1996) and was used in previous studies from our laboratory (Tepper & Seldner, 1999; Tepper et al., 1996). A parent (in most cases, the mother) filled out a food frequency questionnaire for their child, and, to permit child-parent comparisons, an additional questionnaire for their own food intake (data to be reported elsewhere).

Subject characteristics Children's weights and heights were measured in the classroom. Children were dressed in light clothing and without shoes. Weight was measured to the nearest pound and height to the nearest inch using a standiometer. Age of the children was recorded by parents at the time parents completed their child's questionnaire packet. Weight for height percentiles were determined by using age-appropriate growth charts (Frisancho, 1990).

Data analysis Acceptance ratings for the foods were converted to number, (1 ˆ dislike to 5 ˆ like) and analyzed by twoway ANOVA, using PROP status and gender as independent variables. Tukey's LSD test was used to determine group differences in mean liking ratings of the foods. Foods from the food frequency questionnaire (see Appendix) were placed into five major food groups for analysis (grains, vegetables, fruits, dairy, and meats). Because of limitations on a food frequency questionnaire, smaller grouping of foods was not feasible. Portion sizes were determined using the U.S. Food Guide Pyramid for Children (U.S.D.A., 1998). Since the purpose of the study was to examine the relationship between PROP status and overall dietary patterns, the foods were grouped according to their primary ingredients, rather than separating the foods into their component parts (Munoz et al., 1997). For example, pancakes were placed in the grain group, instead of assigning the individual ingredients (fat, egg, milk, flour) to separate groups. Because cheese pizza was consumed in large amounts by the children, this food was considered as a mixed entreÂe and its ingredients were separated into two groups. The cheese contributed to the dairy group and the crust contributed to the grain group. There were no other mixed entreÂe items on the questionnaire that children consumed enough of to be separated in the same manner as cheese pizza.

Since consumption of fat containing foods was of interest, two additional food groups were created. These two groups were sweet and fatty snacks (sodas, cookies, doughnuts, brownies), and discretionary fats (cream cheese, spreads, butter, salad dressings, gravies, etc.). For a complete listing of the foods contained in each group, see the Appendix. Daily servings from each of the food groups were estimated for each child. Differences in reported frequency of consumption from the food groups were analyzed by two-way ANOVA, using PROP status and gender as independent variables. All data were analyzed using SAS (version 6
11 SAS Institute, Cary, NC) for the mainframe at a critical value of p  0
05.

Results Subject characteristics Taste and anthropometric data were collected in 67 children (34 boys, 33 girls). The children were predominantly Caucasian (97%), and were within the 85th percentile for weight for height. 64% were classified as PROP tasters (22 boys, 21 girls) and 36% were nontasters (12 boys, 12 girls). There were no differences in age, weight percentile, or gender as a function of PROP taster status. Food frequencies were not returned for 17 of the children. However, results for the taste tests for the remaining children (N ˆ 50) were not different from the sample as a whole. Therefore, taste data are reported for all 67 children; the dietary data are reported for 50 children. Presently, 25 out of 26 children who were available for follow up testing one year later maintained the same PROP classification in both tests (Spearman's r ˆ 0
92, p  0
001).

Food acceptance ratings Acceptance ratings for the foods are shown in Figure 1. Significant differences between tasters and nontasters were found for raw broccoli and American cheese. Taster children gave lower acceptance ratings to raw broccoli and American cheese than did nontasters [F(1,64) ˆ 4
32, p  0
05 and F(1,61) ˆ 5
86, p  0
05, respectively]. The same trend was also seen for sharp cheese, although the difference was not statistically significant. A significant interaction was also found between gender and PROP taster status for liking of full-fat milk [taster * gender interaction F(1,63) ˆ 5
12, p  0
05], such that taster girls liked full-fat milk less than nontaster girls (p  0
05). The same effect was not seen in boys. No differences in liking between tasters and nontasters were found for the remaining foods.

Genetic taste influences in children Tasters Nontasters

(a) Raw broccoli

*

Cooked broccoli Orange juice Grapefruitorange juice Turkey hotdog Beef hotdog Milk chocolate Semisweet chocolate American cheese

*

Sharp cheese Skim milk (b) Full-fat milk Girls

*

Boys 1

2

3

4

5

Mean liking rating (5 point scale)

Figure 1. (a) Mean acceptance ratings by taster ( ) and nontaster (&) children (SEM) for laboratory test foods. Tasters gave lower acceptance ratings to raw broccoli and American cheese (p  0
05). (b) Mean acceptance ratings by taster and nontaster children (SEM) for full-fat milk and interactions between gender and taster status for acceptance of full-fat milk. Taster girls gave lower acceptance ratings to full-fat milk (p  0
05). There were no differences in acceptability as a function of taster status for the remaining test foods. N ˆ 43 tasters (22 boys, 21 girls) and N ˆ 24 nontasters (12 boys, 12 girls).

Reported consumption of food groups Estimated daily intakes from the food groups are shown in Figure 2. Reported consumption of grains, fruits, vegetables, meats, dairy, and sweet and fatty snacks was not different by taster group. Nontasters reportedly consumed more daily servings of discretionary fats than tasters [F(1,43) ˆ 7
05, p  0
05]. There was an interaction between taster status and gender [F(1,43) ˆ 3
58, p  0
05], such that nontaster girls reportedly consumed 2±3 more servings of discretionary fats per day than either taster girls or boys in both taster groups (p  0
05).

7

Discussion The first objective of this study was to determine if PROP taster status was associated with differences in acceptability of bitter tasting or high-fat foods. Results showed that PROP taster children gave lower acceptance ratings than nontasters to certain bitter foods including raw broccoli and American cheese. Taster girls also gave lower acceptance ratings to full-fat milk than did any of the other groups of children. Although not all studies support the relationship between taste sensitivity to PROP and differences in food acceptance (Mattes & Labov, 1989; Jerzsa-Latta et al., 1990), the present results are most consistent with the findings of Drewnowski et al. (2000) who found lower acceptance of bitter tasting fruits and vegetables in PROPsensitive women and Duffy & Bartoshuk (2000) who reported lower acceptance of high-fat foods among PROP-sensitive adults. Our results complement these earlier findings and suggest that differences in the acceptability of some bitter and high-fat foods may, in part, be genetically-determined, and these differences have their origins in early childhood. The results for the broccoli samples, cheeses and milks can be explained on the basis of the primary sensory attributes of these foods. Raw broccoli contains bitter tasting flavonoids and isothiocyanates that are released during cooking (Betz & Fox, 1994). Thus, cooking might have enhanced the acceptability of broccoli to taster children. The possibility that taster children may be more likely to accept broccoli if it is cooked could have important implications for childhood nutrition and warrants further investigation. Cooking broccoli also softens its texture. Since children generally prefer soft foods to hard foods (Szczesniak, 1990), the softer texture of the cooked broccoli might have also influenced the acceptability of this sample. American cheese contains calcium chloride which is also bitter tasting. Taster children may be more sensitive to bitterness in cheese than nontasters, which could explain the lower acceptance of this food by tasters. Although there was a clear trend for taster children to give lower acceptance ratings to sharp American cheese, as well, this difference was not statistically significant. The reasons for this discrepancy are unknown. Milk-fat provides the characteristic flavour and texture to fluid dairy products. Humans can readily detect differences in fat content in milk based on creaminess and viscosity; flavour plays a less critical role in the perception of fat in this medium (Mela, 1988; Drewnowski et al., 1989; Richardson et al., 1993; Tepper & Kuang, 1996). Duffy et al. (1996, abstract) previously demonstrated that PROP tasters perceived

8

K. L. Keller et al. 6 (a)

Tasters

(b)

Nontasters

5

+

Servings/d

4 * 3 2 1 0

Grains

Fruits

Veg.

Meats

Dairy

Sweet-fats All

Discretionary fats Girls Boys

Figure 2. (a) Mean reported daily intakes (SEM) of tasters ( ) and nontasters (&) for food groups from the food frequency questionnaire. Reported consumption of grains, fruits, vegetables, meats, dairy, and sweet and fatty snacks did not differ as a function of taster status. (b) Mean reported intake (SEM) for tasters and nontasters and interactions between gender and taster status for reported daily intake of discretionary fats. N ˆ 30 tasters (14 girls, 16 boys); N ˆ 20 nontasters (8 girls, 11 boys). *Reported daily intake of discretionary fats was higher for nontasters than tasters (p  0
05). + Nontaster girls reportedly consumed more daily servings of discretionary fats than taster girls, and boys in both groups (p  0
05).

more creaminess in high-fat milk samples than nontasters. Evidence suggests that this variation may be due to anatomical differences between tasters and nontasters (Bartoshuk et al., 1994; Tepper & Nurse, 1997). In contrast to Duffy et al.'s (1996) findings, Drewnowski et al. (1998) found no differences among taster groups in the perception of creaminess in sucrose-sweetened dairy products. It is possible that dairy products are not perceived differently by tasters and nontasters of PROP when they are sweetened. In addition, methodological differences between studies, such as the use of less sensitive scaling techniques, might have led to the observed disparity in results (See Prutkin et al., 2000). Milk also contains calcium chloride, but at relatively low concentrations (Muldoon, 1970). Calcium salts impart bitter/astringent notes in milk that are most noticeable in skim milk (Forss, 1967). Therefore, if bitterness was a salient characteristic for the children's acceptance of milk in our study, then tasters would have given lower ratings to skim milk than did nontasters. However, the present study found taster-group differences in the acceptability of full-fat milk but not fat-free milk. These data suggest that differences in fat content, not bitterness, influenced the acceptance of milk in our study. The present interpretations are based on the predominant sensory characteristics of these foods. It is recognized, of course, that all foods are mixtures of taste, olfactory, and textural properties. For example, in addition to having bitter taste properties, American

cheese is also a high-fat food, a property that might also have influenced ratings between tasters and nontasters. As noted previously, cooked broccoli has markedly different flavour and texture properties than raw. As in most taste studies, even in adults, it is difficult to identify the particular characteristic that drives liking (Lawless & Heymann, 1998). It is assumed that the characteristics varied in these samples were among the most important, but we cannot be certain. We intentionally selected foods with only moderatelystrong taste qualities. This was done to avoid the possibility that the children would reject very strongtasting foods and refuse to complete the protocol. However, several foods (e.g. chocolates, hotdogs, juices) did not show the expected differences between tasters and nontasters. It is possible that taster status is not related to the acceptance of these foods. It is also conceivable that at this young age, children may not notice modest variations in the sensory characteristics of some foods. Young children are less discriminating than adults (Zandstra & De Graaf, 1998) and differences in flavour intensity of the taste stimuli might not have been noticeable to children in this study. Consequently, small differences in liking of the foods reported by tasters and nontasters might not have been detectable. This explanation is supported by subsequent data from our laboratory showing that taster children gave lower acceptance ratings to grapefruit-orange juice when the ratio of grapefruit to orange juice in the blend was increased (p  0
05; Keller & Tepper, unpublished).

Genetic taste influences in children

Since children become more discriminating of sensory differences in foods with increasing age (Zandstra & De Graaf, 1998), it would be interesting to re-test these children when they are older to determine whether such differences appear later in development. The present study supports the earlier work of Anliker et al. (1991) who also found taster-group differences in food acceptance among 5±7-years old children. However, several findings in this earlier study vary from those in the present study. Anliker et al. reported that nontaster children preferred cheese (in agreement with present results) whereas tasters preferred milk (in contrast to present results). They used a rank-order preference design in which children selected foods oneat-a-time from the most liked to the least liked, and interpreted their findings based on the perceived bitterness of these two foods. On this basis it seems logical that nontaster children might select cheese first, whereas taster children might select milk first. Ranking is a forced-choice procedure that requires subjects to judge foods against each other. In contrast, children in the present study gave independent ratings to each food. Thus, each food was judged for its unique sensory characteristics and not compared to other foods. This difference in methodology may explain the disparity in results. Another difference between the present study and the work of Anliker et al. (1991) is that the earlier study found no taster-group differences in liking for raw broccoli. As discussed previously, some studies in adults have shown lower acceptance of cruciferous vegetables by tasters (Drewnowski et al., 2000; Kaminski et al., 2000), but this finding is not universal (Jerzsa-Latta et al., 1990; Mattes & Labov, 1989; Duffy & Bartoshuk, 2000). Many factors mediate the acceptance or rejection of vegetables in addition to taste. These factors include familiarity, social desirability and cultural norms (Resnicow et al., 1997). The relative influence of these factors on the reported liking of vegetables has yet to be determined. Finally, the children studied by Anliker et al. (1991) were slightly older than those participating in this study, and this age difference might also have contributed to the differences between the two studies. The second objective of this study was to determine whether PROP taster status influenced the dietary intakes of the children. Results showed that estimated daily intakes from most of the major food groups did not vary by taster status. However, nontaster children reportedly consumed more daily servings of discretionary fats than tasters. These differences were primarily due to nontaster girls who reportedly consumed 2±3 more daily servings of discretionary fats than did the other groups of children. To our knowledge, these

9

data provide the first published evidence in children suggesting that the inherited taste response to PROP is associated with differences in reported consumption of discretionary fats. These results are intriguing in light of our previous findings in adults, showing that PROP nontasters preferred high-fat to low-fat salad dressings in laboratory taste tests (Tepper & Nurse, 1998) and another study in adults showing that those with lower taste sensitivity to PROP had an increased acceptance for a range of fat-containing foods (Duffy & Bartoshuk, 2000). Since salad dressing is a type of discretionary fat, together these findings could suggest that PROP nontasters show greater acceptance of this class of fats. We found no differences in the reported consumption of sweet and fatty snacks as a function of taster status, a finding that would have carried obvious implications for childhood nutrition. The relationship between PROP and acceptance of sweet-fat foods is a controversy in the literature and warrants further investigation. Previous studies using this screening method in adults identified approximately 60% tasters and 40% nontasters (Lawless, 1980; Mela, 1990). This is the first study to use this method to classify taster status in young children. Because test-retest reliability was high, and the relative proportion of tasters and nontasters was the same as those obtained with this same method in adults, it is thought that this method can convincingly divide children into tasters and nontasters. We note, however, that thresholds are a more definitive method for separating tasters from nontasters (Prutkin et al., 2000). Thresholds were not used in the present study, as we were not confident that children this young could complete the task. Studies using thresholds in children have reported the percentage of nontasters as varying from approximately 18±35% (Whissel-Buechy, 1990; Harris & Kalmus, 1949; Anliker et al., 1991). While present methods may be slightly less sensitive than thresholds, the percentage of nontasters identified in the present work was within this range. As an additional explanation according to Whissel-Buechy (1990), variability in the reported percentages of tasters and nontasters might be due to variation in the gene frequency of PROP in different population groups. This might provide additional explanation for differences found in the distribution of tasters and nontasters across studies. Although adult females are somewhat more sensitive to PROP than males, no consistent gender differences are reported in children under the age of 10 years (Whissel-Buechey, 1990; Glanville et al., 1964) which was also observed in the present study. Thus, gender differences in PROP taster status do not appear to explain why PROP-related differences in acceptance of

10

K. L. Keller et al.

full-fat milk and reported consumption of discretionary fats was limited to girls. Experiential factors might have played a role in these results. For example, Birch and colleagues (Johnson & Birch, 1994; Cutting et al., 1999; Fisher & Birch, 1999) have suggested that parent's concerns about their own overweight or the fear that their child may become overweight can influence their child-feeding practices. However, attempts by parents to strictly control their children's food intake increased children's preferences for and selection of high-fat snack foods, instead of decreased them (Johnson & Birch, 1994). Thus, the frequency of exposure to high-fat foods or the social context in which this exposure takes place could mediate the role of PROP status in the expression of fat preference in childhood. Restrictive strategies are more likely to adversely affect eating behaviours in girls (Cutting et al., 1999). Since girls are also at greater risk for excess weight gain and eating disorders (Dietz, 1998) at increasingly younger ages, further investigation of this relationship in girls seems warranted. This study had several limitations. First, collecting nutritional information in young children is problematic. Consequently, there is no accepted standard for dietary reporting for this age group (Carter et al., 1981; Stuff et al., 1983; Brownell & Foreyt, 1986; Baranowski et al., 1986). Since total food consumption is low in young children, the use of a food frequency questionnaire presumably allowed us to capture intake from a greater variety of foods over a longer time frame than other methods would have permitted (Block, 1982). Second, many foods on a frequency questionnaire are reported by groups rather than singly, and these groupings are based on nutritional content (e.g. citrus fruits), not taste (e.g. grapefruit vs. oranges). This is a typical problem in research that attempts to link taste to eating behaviour and consequently, this limitation made it difficult to relate the food acceptance ratings given in the laboratory to dietary intakes of those same foods or related foods. Grouping the foods by similar taste qualities (e.g. bitter tasting vegetables, sweet-fat dairy desserts) was also attempted. However, reported intake from these groups was low, and group differences were not detectable. Direct observations of food intake in the classroom setting may offer a more feasible approach to addressing this question. Finally, the children in this study came from a homogeneous sample of middle-upper income families, and probably had an above average exposure to a variety of foods at the preschool. Whether these results will generalize to children in the population at-large remains to be tested. In conclusion, these results suggest that genetic taste factors may play a more important role in the acquisition of food preferences in childhood than previously

realized. Moreover, PROP status may serve as a convenient marker for understanding the development of food preferences across maturational periods, and consequently, following children over time might help clarify how genetic and environmental variables interact to influence adult food preferences and dietary selection. Although all the results did not support out initial predictions, this study provides important insight into how genetic sensitivity to PROP influences young children's food selection. A better understanding of the genetic basis of taste may provide novel approaches to solving childhood feeding problems and ultimately lead to the design of better dietary prevention strategies for children.

Acknowledgements The studies were carried out at the Rutgers University Nutritional Sciences Preschool. The work was conducted in partial fulfillment of a doctoral thesis by KLK and was presented at the SSIB annual meeting in Dublin, Ireland, July 2000. The authors thank Harriet Worobey, the preschool director, and preschool teachers Pam Mancini and Donna Kuchinski for help in carrying out this study, as well as the many parents and children who participated.

References Anliker, J. A., Bartoshuk, L. M., Ferris, A. M. & Hooks, L. D. (1991). Children's food preferences and genetic sensitivity to the bitter taste of 6-n-propylthiouracil (PROP). American Journal of Clinical Nutrition 54, 316±320. Baranowski, T., Dworkin, T., Henske, J. C., Clearman, D. R., Dunn., J. K., Nader, P. R. & Hooks, P. C. (1986). The accuracy of children's self-reports of diet: Family health project. Journal of the American Dietetic Association 86, 1382±1385. Bartoshuk, L. M. (1979). Bitter taste of saccharin related to the genetic ability to taste the bitter substance 6-npropylthiouracil (PROP). Science 203, 934±935. Bartoshuk, L. M., Duffy, V. B. & Miller, I. J. (1994). PTC/ PROP tasting: anatomy, psychophysics, and sex effects. Physiology and Behavior 56, 1165±1171. Bartoshuk, L. M., Rifkin, B., Marks, L. E. & Hooper, J. E. (1988). Bitterness of KCl and benzoate: related to genetic status for sensitivity to PTC/PROP. Chemical Senses 13, 517±528. Betz, J. M. & Fox, W. D. (1994). High-performance liquid chromatographic determination of glucosinolates in Brassica vegetables. American Chemical Society Symposium Series. American Chemical Society 546, 181±196. Birch, L. L. (1979). Dimensions of preschool children's food preferences. Journal of Nutrition Education 11, 77±80. Birch, L. L. (1989). Effects of experience on the modification of food acceptance patterns. Annals of the New York Academy of Sciences 561, 209±216.

Genetic taste influences in children 11 Birch, L. L. & Marlin, D. W. (1982). I don't like it; I never tried it: effects of exposure on two-year-old children's food preferences. Appetite 3, 353±360. Birch, L. L., McPhee, L., Schoba, B. C., Pirok, E. & Steinberg, L. (1987). What kind of exposure reduces children's food neophobia? Appetite 9, 171±178. Birch, L. L., McPhee, L., Steinberg, L. & Sullivan, S. (1990). Conditioned flavor preferences in young children. Physiological Behaviour 47, 501±505. Birch, L. L., Zimmerman, S. I. & Hind, H. (1980). The influence of social-affective context on the formation of children's food preferences. Child Development 51, 856±861. Blakeslee, A. F. (1931). Genetics of sensory thresholds: taste for phenylthiocarbamide. Proceedings of the National Academy of Sciences 18, 120±130. Block, G. (1982). A review of validations of dietary assessment methods. American Journal of Epidemiology 115, 492±505. Brownell, K. D. & Foreyt, J. P. (1986). Handbook of eating disorders: physiology, psychology, and treatment of obesity, anorexia, and bulimia. New York: Basic Books. Burt, J. V. & Hertzler, A. A. (1978). Parental influence on the child's food preferences. Journal of Nutrition Education 10, 127±128. Carter, R. L., Sharbaugh, C. O. & Stapell, C. A. (1981). Analysis of data from a pediatric population: reliability and validity of the 24-hour recall. Journal of the American Dietetic Association 79, 542±547. Chen, A. W. & Resurreccion, A. V. (1996). Age appropriate hedonic scales to measure food preferences of young children. Journal of Sensory Studies 11, 141±163. Cowart, B. J. (1981). Development of taste perception in humans: sensitivity and preference throughout the life span. Psychology Bulletin 90, 43±73. Cutting, T. M., Fisher, J. O., Grimm-Thomas, K. & Birch, L. L. (1999). Like mother, like daughter: familial patterns of overweight are mediated by mothers' dietary disinhibition. American Journal of Clinical Nutrition 69, 608±613. Dietz, W. H. (1998). Childhood Obesity. In Shils, M. E., Olson, J. A., Shike, M. & Ross, A. C. (Eds.), Modern nutrition in health and disease (8th Ed.), pp. 1071±1079. Williams & Wilkins: New York. Drewnowski, A., Henderson, S. A. & Shore, A. F. (1997). Taste responses to naringin, a flavonoid, and the acceptance of grapefruit juice are related to genetic sensitivity to 6-n-propylthiouracil. American Journal of Clinical Nutrition 66, 391±397. Drewnowski, A., Henderson, S. A. & Barratt-Fornell, A. (1998). Genetic sensitivity to 6-n-propylthiouracil and sensory responses to sugar and fat mixtures. Physiology and Behavior 63, 771±777. Drewnowski, A., Henderson, S. A., Hann, C. S., Berg, W. A. & Ruffin, M. T. (2000). Genetic taste markers and preferences for vegetables and fruit of female breast care patients. Journal of the American Dietetic Association 100, 191±197. Drewnowski, A., Shrager, E. E., Lipsky, C., Stellar, E. & Greenwood, M. R. C. (1989). Sugar and fat: sensory and hedonic evaluation of liquid and solid foods. Physiology and Behavior 45, 177±183. Duffy, V. B. & Bartoshuk, L. M. (2000). Food acceptance and genetic variation in taste. Journal of the American Dietetic Association 100, 647±655.

Duffy, V. B., Bartoshuk, L. M., Lucchina, L. A., Snyder, D. J. & Tym, A. (1996). Supertasters of PROP (6-npropylthiouracil) rate the highest creaminess to high-fat milk products. Chemical Senses 21, abstract. Fisher, J. O. & Birch, L. L. (1999). Restricting access to palatable foods affects children's behavioral response, food selection and intake. American Journal of Clinical Nutrition 69, 1264±1272. Forss, D. A. (1967). Flavors of dairy products: a review of recent advances. Journal of Dairy Science 52, 832±840. Frisancho, A. R. (1990). Anthropometric standards for the assessment of growth and nutritional status. Ann Arbor, MI: University of Michigan Press. Glanville, E. V. & Kaplan, A. R. (1965). Food preferences and sensitivity of taste for bitter compounds. Nature 205, 851±853. Glanville, E. V., Kaplan, A. R. & Fischer, R. (1964). Age, sex, and taste sensitivity. Journal of Gerontology 14, 474±478. Harris, H. & Kalmus H. (1949). The measurement of taste sensitivity to phenylthiourea (PTC). Annals of Eugenics 15, 24±31. Jerzsa-Latta, M., Krondl, M. & Coleman, P. (1990). Use and perceived attitudes of cruciferous vegetables in terms of genetically-mediated taste sensitivity. Appetite 15, 127±134. Johnson, S. L. & Birch, L. L. (1994). Parents' and children's adiposity and eating style. Pediatrics 94, 653±661. Kaminski, L. C., Henderson, S. A. & Drewnowski, A. (2000). Young women's food preferences and taste responsiveness to 6-n-propylthiouracil (PROP). Physiology and Behavior 68, 691±697. Kimmel, S. A., Sigman-Grant, M., Guinard., J. Sensory testing with young children. Food Technology 48, 90±99. Koivisto, U. K. & Sjoden, P. O. (1996). Reasons for rejection of food items in Swedish families with children aged 2±17. Appetite 26, 89±103. Krebs-Smith, S. M., Cook, A., Subar, A. F., Cleveland, L., Friday, J. & Kahle, L. L. (1996). Fruit and vegetable intakes of children and adolescents in the United States. Archives of Pediatric and Adolescent Medicine 150, 181±186. Lawless, H. T. (1980). A comparison of different methods used to assess sensitivity to the taste of phenylthiocarbamide (PTC). Chemical Senses 5, 247±255. Lawless, H. T. & Heymann, H. (1998). Sensory evaluation of food: principles and practices. New York: Chapman & Hall. Looy, H. & Weingarten, H. P. (1992). Facial expressions and genetic sensitivity to 6-n-propylthiouracil predict hedonic response to sweet. Physiology and Behavior 52, 75±82. Mattes, R. C. & Labov, J. (1989). Bitter taste responses to phenylthiocarbamide are not related to dietary goitrogen intake in human beings. Journal of the American Dietetic Association 89, 692±694. Mela, D. J. (1988). Sensory assessment of fat content in fluid dairy products. Appetite 10, 37±44. Mela, D. J. (1990). Gustatory perception of isohumulones: influence of sex and thiourea taster status. Chemical Senses 15, 485±490. Muldoon, P. J. (1970). The determination of calcium and chloride ion concentrations in milk using ion selective electrodes. Ann Arbor Michigan: University Microfilms x, 76 leaves.

12

K. L. Keller et al.

Munoz, K. A., Krebs-Smith, S. M., Ballard-Barbush, R. & Cleveland, L. E. (1997). Food intakes of US children and adolescents compared with recommendations. Pediatrics 100, 323±329. Prutkin, J., Duffy, V. B., Etter, L., Fast, K., Gardner, E., Lucchina, L. A., Snyder, D. J., Tie, K., Weiffenbach, J. & Bartoshuk, L. M. (2000). Genetic variation and inferences about perceived taste intensity in mice and men. Physiology & Behavior 69, 161±173. Resnicow, K., Davis-Hearn, M., Smith, M., Baranowski, T., Lin, L. S. & Baranowski, J. (1997). Social-cognitive predictors of fruit and vegetable intake in children. Health Psychology 16, 272±276. Richardson, N. J., Booth, D. A. & Stanley, N. L. (1993). Effect of homogenization and fat content on oral perception of low and high viscosity model creams. Journal of Sensory Studies 8, 133±143. Rozin, P. & Vollmecke, T. A. (1986). Food likes and dislikes. Annual Review of Nutrition 6, 433±456. Snyder, L. H. (1931). Inherited taste deficiency. Science 74, 151±152. Stuff, J. E., Garza, C., Smith, E. O., Nichols, B. L. & Montandon, C. M. (1983). A comparison of dietary methods in nutritional studies. American Journal of Clinical Nutrition 37, 300±306. Szczesniak, A. S. (1990). Texture: is it still an overlooked food attribute? Food Technology 44, 86, 88, 90, 92, 95. Tepper, B. J. (1998). 6-n-propylthiouracil: a genetic marker for taste, with implications for food preference and

dietary habits. American Journal of Human Genetics 63, 1271±1276. Tepper, B. J. & Kuang, T. (1995). Perception of fat in a milk model system using multidimensional scaling. Journal of Sensory Studies 11, 175±190. Tepper, B. J., Trail, A. C. & Shaffer, S. E. (1996). Diet and physical activity in restrained eaters. Appetite 27, 51±64. Tepper, B. J. & Nurse, R. J. (1997). Fat perception is related to PROP taster status. Physiology and Behavior 61, 949±954. Tepper, B. J. & Nurse, R. J. (1998). PROP taster status is related to fat perception and preference. Annals of the New York Academy of Sciences 855, 802±804. Tepper, B. J. & Seldner, A. C. (1999). Sweet taste and intake of sweet foods in normal pregnancy and pregnancy complicated by gestational diabetes mellitus. American Journal of Clinical Nutrition 70, 277±284. U.S. Department of Agriculture. (1999) Food guide pyramid for children. http.//www.usda.gov/cnpp/KidsPyra/ KIDPYR.bw.pdf. Whissel-Buechy, D. (1990). Effects of age and sex on taste sensitivity to phenylthiocarbamide (PTC) in the Berkeley Guidance sample. Chemical Senses 15, 39±57. Zandstra, E. H. & De Graaf, C. (1998). Sensory perception and pleasantness of orange beverages from childhood to old age. Food Quality & Preference 9, 5±12.

Appendix: Foods contained in food groups from food frequency questionnaire Food groups

Foods contained in each group

Grains

white bread, wheat bread, crackers, low-crackers, tortillas, muffins, English muffins, bagels, pita bread, pancakes, waffles, toaster pastries, hot cereal, cold cereal, granola and granola bars, white rice, brown rice, pasta, cheese pizza (crust)

Fruits

fresh fruit, canned fruit, fruit juice

Vegetables

lettuce, dark green leafy vegetables (broccoli, spinach, kale, etc.), cabbage, carrots, celery, tomatoes, tomato sauce, ketchup, green beans, corn, peas, white potatoes, French fries, sweet potatoes, zucchini, squash

Dairy

whole milk, reduced fat milk, skim milk, chocolate milk, reduced fat chocolate milk, yogurt, low fat yogurt, no fat yogurt, cheese, reduced fat cheese, no fat cheese, cottage cheese, reduced fat cottage cheese, pudding, reduced calorie pudding, fat free pudding, ice cream, rich ice cream, reduced fat ice cream, fat free ice cream, frozen yogurt, milkshakes

Meats

nuts, seeds, peanut butter, nut butters, tofu, vegetarian meat, bacon, beef (roast, steak, ground meat), chicken, turkey, duck, pork, veal, lamb, canned fish in oil, canned fish in water, fresh fish, eggs, egg substitutes, lunch meat, reduced fat lunch meat, hot dogs, hamburgers, sausage, Italian sausage

Discretionary fats

olive oil, vegetable oil, vegetable shortening, lard, margarine, light margarine, butter, mayonnaise, no cholesterol mayonnaise, no fat mayonnaise, sour cream, reduced fat sour cream, salad dressing, reduced calorie salad dressing, fat free salad dressing, cream cheese, reduced fat cream cheese, cream, half and half, non-dairy creamer, gravy

Sweets/snacks

sugar, artificial sweetener, soda, cookies, reduced fat cookies, brownies, doughnuts, pastries, reduced fat pastries, cake, cake with icing, pie, candy, hard candy, potato chips