Flavor enhancement as a tool for increasing pleasantness and intakeof a snack product among the elderly
Research Report Flavor enhancement as a tool for increasing pleasantness and intake of a snack product among the elderly S. Koskinen * ,N.Ka ¨ lvia ¨ inen 1 , H. Tuorila Department of Food Technology, University of Helsinki, P.O. Box 27, Helsinki 00014, Finland Accepted 29 April 2003 Abstract A yogurt-like fermented oat bran product, flavored with regular and heightened concentrations of red currant aroma, was tested in two tasting sessions (side-by-side) and, between these, in a six-day home-use (monadic testing daily, 3 þ 3 packages of the snack) by the elderly (n ¼ 50; mean age 73.7, range 63–85 years) and the young (n ¼ 58; mean age 23.1, range 18–34 years). The subjects rated the odor and flavor intensity and pleasantness and also conducted an odor detection and identification test. In home-use, the subjects reported the quantity consumed, willingness to eat, buy or recommend the snack. The young outperformed the elderly in the olfactory test. The heightened aroma samples were initially rated as less pleasant by both age groups, but among the elderly, the ratings given to the two samples merged during exposure. For the young, the large difference in perceived odor and flavor intensities reflected marked differences in pleasantness, while the elderly were less responsive to intensity differences in their pleasantness ratings. Overall, both age groups ate less of the heightened aroma sample. Despite the impaired olfactory capabilities of the elderly, no clear indication of benefit of the enhanced flavor was found for either pleasantness or intake. q 2003 Elsevier Ltd. All rights reserved. Keywords: Elderly; Flavor enhancement; Olfactory test Introduction People’s food choices change during the course of aging. The daily volume of consumed foods and beverages (grams/day) declines as a function of age. Adjustments in intake are due to many physiological changes associated with age, including altered function of smell and taste (Drewnowski & Shultz, 2001). According to Rolls (1992), the elderly rate flavor perception as a strong determinant of their food choices. However, the elderly are often unaware of their impaired ability to smell and report no decrease in their appreciation of foods (Wysocki & Pelchat, 1993). Rolls (1993) suggests that when sensory losses occur gradually over a period of years, their impact on appetite also gradually declines. In the geriatric literature, the term successful aging has been brought up, and it can refer to any type of function (Rowe & Kahn, 1987). The term healthy aging is used in a similar context. Simons, McCallum, Friedlander, and Simons (2000) suggest that healthy aging is associated with reduced and delayed disability. Hoffman, Ishii, and Macturk (1998) reported that individuals’ overall health status, other sensory impairments, and functional limitations were associated with an increase in the rate of chemosensory disorders. According to Schiffman and Warwick (1993), addition of flavor can compensate for age-related perceptual losses, improve food palatability and acceptance, and increase intake in the elderly. Tuorila, Niskanen, and Maunuksela (2001) found a slight indication that by adding an appropriate aroma the hedonic quality of food could be improved among the elderly. Mattes and Cowart (1994) proposed that while increasing the appeal of foods by adding flavor may enhance the quality of life for people who have suffered chemosensory loss, it would probably have little nutritional relevance. Griep, Mets, and Massart (2000) found a clear increase in preference for flavor-enhanced food in the elderly, but a preference for a particular food 0195-6663/03/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0195-6663(03)00054-0 Appetite 41 (2003) 87–96 www.elsevier.com/locate/appet 1 Present address: Seina ¨ joki Polytechnic, Finland. * Corresponding author. E-mail address: sari.ml.koskinen@helsinki.fi (S. Koskinen). did not necessarily imply increased intake. De Jong, De Graaf, and van Staveren (1996) studied different sucrose concentrations in breakfast items and showed that flavor concentration rated as being most pleasant by the elderly in a laboratory setting failed to predict pleasantness and food intake in realistic settings. Wysocki and Pelchat (1993) suggest that olfactory loss is not uniform across odorants. Additionally, De Graaf, Polet, and van Staveren (1994) and De Graaf, van Staveren, and Burema (1996) reported that differences were present between age groups in pleasantness ratings of different types of flavors. According to Murphy (1993), the age effects on preference do not appear to be related to generally lower perceived intensity. The overall odor or flavor complex may be differentially perceived by the elderly and the young. Thus, when some of the components fall below an older person’s threshold, the altered flavor may then drive the preference. Most preference tests are based on first impressions of an evaluated product, and thus do not reflect preferences that result after a longer exposure to the product. The natural eating environment at home also contributes to realistic hedonic ratings (Porcherot & Issanchou, 1998). Moreover, Cardello and Maller (1982) have questioned the respondent’s ability to form an opinion on the basis of a very short exposure to a product that is normally used in larger quantities. Sensory impairments over the course of aging have been widely studied, and relatively abundant literature also exists on the hedonic responses of the elderly towards foods with enhanced flavor. However, little is known about the effects of exposure on the hedonic responses of the elderly towards such foods. In addition, very few reported studies are available that examine the relation between olfactory capability and hedonic responses over time. One such study has been conducted by Pelchat and Schaefer (2000), who hypothesized that the age difference in liking a vanilla beverage might be mediated by different olfactory sensitivity. However, they found no correlation between odor sensitivity (measured with phenyl ethyl alcohol, PEA) and liking of the vanilla beverage at any phase of the study. The aim of the present study was to examine the effects of heightened aroma concentration on the pleasantness and intake of a snack product in elderly and young consumers over a longer exposure. To obtain additional information on the effects of olfactory sensitivity and health-related issues on sensory and hedonic responses, the elderly were sub- grouped based on their olfactory function and health status. We hypothesized that the enhancement of flavor by adding aroma would increase the pleasantness and intake of the elderly with diminished olfactory capability (De Graaf et al., 1994, 1996). Furthermore, we anticipated that poor health status of the elderly would impair their olfactory perform- ance (Schiffman, 1997) and predict decreased intake (Rolls, 1993; Schiffman, 1994). Methods Subjects A total of 57 elderly and 62 young subjects were initially recruited, but seven of the former and four of the latter group missed one of the three test sessions and were therefore excluded from the data analysis. The indepen- dently living elderly subjects (n ¼ 50; 38 females and 12 males, mean age 73.7, range 63–85 years) were recruited from the visitors of four service centers for the elderly in Helsinki by personal contacts. The young subjects (n ¼ 58; 55 females and 3 males, mean age 23.1, range 18–34 years) were recruited from the staff and students of the University of Helsinki by announcements. The study was conducted according to the general principles of sensory testing in our laboratory, which have been approved by the ethical committee of the faculty. The subjects signed an informed consent before the study began. The subjects participated in two evaluation sessions lasting approximately 30 min each. In addition to the tasting sessions, the subjects participated in a six-day home-use test. Subjects received a small food reward after the last evaluation session. Samples The sample material was a fermented oat bran product (Yosa w , Bioferme, Finland), which is a cereal-based snack food similar to flavored yogurt. Samples were flavored with a 20% (w/w) red berry jam (Valio Ltd, Finland), containing strawberry, red currant, lingonberry, and chokeberry. Two concentrations of red currant aroma (Valio Ltd, Finland) were also added to the samples. The aroma concentrations (regular 0.02% and heightened 0.78%) were selected after pre-testing by an experienced laboratory panel (n ¼ 10; 7 females, 3 males, aged 25–41 years), which evaluated the odor (by sniffing) and flavor intensity (by tasting) of five concentrations of red currant aroma in the oat bran product. The two selected concentrations differed from each other significantly in odor and flavor intensities (means: 3.8 vs. 7.1 for odor and 3.5 vs. 7.4 for flavor on a 9-point scale). A sensory profile of the two snack product samples was created by a laboratory panel (n ¼ 9; 7 females, 2 males, aged 25 –41 years) trained to evaluate attributes selected in advance. The attributes were evaluated on a 9-point scale that was anchored from 1 ¼ weak to 9 ¼ strong. The heightened aroma sample was perceived as more intense in currant odor and total odor, but less intense in strawberry odor. It was also perceived to be less sweet, more bitter and more intense in currant flavor and in overall flavor, and stronger in after-taste than the sample with a regular aroma concentration. The heightened aroma concentration caused a slight off-flavor described as ‘artificial’. The manufacturer packaged the samples in 150 ml white plastic cups, which were coated with yellow cardboard and sealed with white aluminum lids. The packaging was similar S. Koskinen et al. / Appetite 41 (2003) 87–9688 to the commercial version, except for being blank. The samples were stored refrigerated at 4 8C. Tasting sessions The tasting sessions for the young subjects were conducted in the sensory laboratory at the University of Helsinki, and for the elderly in the dining rooms or clubrooms of the service centers. Tests were conducted before and after a six-day home-use period. The subjects received two samples of the snack product successively in both sessions and were asked to rate the odor intensity (by sniffing) and flavor intensity (by tasting) of the samples using a 9-point intensity scale that was anchored from 1 ¼ weak to 9 ¼ strong. Overall pleasantness of the samples was rated on a 9-point hedonic scale, anchored from 1 ¼ do not like at all to 9 ¼ like very much. The odor intensity was rated right after opening the lid, and the pleasantness and flavor intensity after tasting 3–4 spoonfuls of the sample. The serving temperature of the samples was 4–6 8C. Subjects rinsed their mouths with tap water before tasting the samples and between samples. The samples were coded with three-digit random numbers, and their presen- tation order was individually randomized for each subject. Before evaluating the samples in the first tasting session, the subjects filled in a background questionnaire covering their demographic background and nutritionally related diseases (Table 1). Furthermore, the subjects reported their familiarity and liking of the target product, as well as those of yogurt and porridge. Since no main or interaction effects of those variables were found on any of the ratings, the data on familiarity and liking are not reported here. The lack of effect of familiarity and liking of the target product on the ratings was probably due to its resemblance to yogurt. Home-use test For the home-use test, subjects received three cups (150 g each) of the snack product at regular aroma concentration and three cups at heightened aroma concen- tration and were instructed to refrigerate them. The samples were coded with numbers 1–6, indicating the eating order of the samples. Half of the subjects ate the regular aroma samples first and half ate the heightened aroma samples first. The subjects were asked to rate one sample a day at any time of the day they wanted. Similarly to the tasting sessions, the odor intensity was rated right after opening the lid, and the pleasantness and flavor intensity after tasting 3–4 spoonfuls of the sample. The scales used were the same as in the tasting sessions. Subjects were also asked to evaluate the amount of eaten snack product using a 5-point visual scale with simple boxes black and white in color, the black part indicating the amount of left-over in the cup after eating. In data treatment, the visual scale points were converted into rough percentages: 0 (sample tasted only), 25, 50, 75 and 100% consumed of the total amount. The subjects were also asked to mark the time and meal context in which they ate it. The data on the eating context and time are not reported here. After the third and sixth days, subjects rated separately their boredom with and willingness to eat and buy the product and recommend it to their friends on a 9-point scale, anchored from 1 ¼ not willing/bored at all to 9 ¼ very willing/bored. Originally the subjects (young and elderly separated) were divided into three groups, which were given different information on the samples for the home-use. One group in both age cohorts was informed about the organic quality of the sample and another group received information about the sample’s healthiness and low fat content. A third group served as control, receiving no additional information, except for the list of ingredients. The written information was provided on the front page of the home-use booklet. Preliminary data analysis revealed no significant main effects and only one significant interaction of day and aroma on the pleasantness ratings of the young, implicating an increase in pleasantness of the heightened aroma sample in the presence of health-related information. Since the effects of information were negligible, and not relevant enough for this study, the data were pooled together over the information conditions. Olfactory test As the final task in the second tasting session, most of the subjects (53 young, 39 elderly) participated in a test that measured their olfactory performance, the European Test of Olfactory Capabilities (ETOC). The test consists of 16 common food and non-food odors that were presented to the subjects in 16 series of four vials: three of the vials were blanks and one contained the odor stimulus. Subjects were asked first to detect which of the four vials contained the odor (detection). After detection, the subjects were Table 1 Total number, gender distribution, age distribution, mean age, and nutritionally related diseases of the subjects Age group Characteristics Total number Gender (male/female) Age range (years) Mean age (years) High blood pressure ðnÞ High cholesterol ðnÞ Diabetes ðnÞ No diseases ðnÞ Young 58 3/55 18– 34 23.1 2 0 0 56 Elderly 50 12/38 63– 85 73.7 20 13 7 20 S. Koskinen et al. / Appetite 41 (2003) 87–96 89 presented a list of four possible odors and asked to select the one they thought was correct (identification) (for more information on the test, see (Thomas-Danguin et al., 2003a)). Data analysis Repeated measures analysis of variance was used to test the within-subject effect of aroma concentration of the sample (two levels) and the day of eating (three levels) on sensory and hedonic ratings and intake ratings in home-use. Age group (two levels) was used as a between-subject factor. The boredom ratings and the willingness ratings to eat, buy, and recommend were tested using aroma (two levels) and day (two levels, third and sixth day) as within-subject factors and age group (two levels) as a between-subject factor. In the tasting sessions, the within-subject effects tested were aroma concentration of the sample (two levels) and the session (two levels, before and after the home-use). Age group (two levels) was used as a between-subject factor. These statistical analyses were also conducted with gender as a between- factor to examine its effect. No trend towards the main effect or interaction of gender was found, and therefore, the effect of gender was not treated in results. The data of the olfactory test (ETOC) were analyzed with one-way analysis of variance, comparing the detection and identification scores of the young and the elderly. The effect of gender was examined in the group of the elderly with one-way analysis of variance. The olfactory scores were compared with the mean scores of age-matched sub-groups in data on 1100 subjects using paired t-test. Furthermore, the elderly were divided into two subgroups based on their identification scores in ETOC using a score of 10 as a cut-off point: (subgroups: identification score 0 –10 and 11 –16). With this cut-off point we could expect differences in perceived odor and flavor intensities (Tuorila & Koskinen, 2003). These two subgroups were referred to as ‘low score’ and ‘high score’, respectively. Next, the effect of perform- ance in the odor identification test on hedonic ratings and intake was tested using repeated measures analysis of variance (within aroma, between identification score). The odor identification scores of the young were not subdivided because of their low variation. The elderly were also divided into two subgroups based on the number of nutritionally related diseases they had. These subgroups were designated ‘good health’ (one or no diseases) and ‘poor health’ (two or more diseases). The difference between these groups in odor detection and identification scores and intake was tested with one-way analysis of variance. The variation in the within-groups degrees of freedom associated with F-values is due to some missing values in the data of the elderly. In the case of intake, there was also missing data from the young. The number of subjects used in analyses of the intakes was 53 (out of 58) for the young and 42 (out of 50) for the elderly. Results Tasting sessions Sensory ratings Both age groups rated odor intensity as stronger in the heightened aroma sample (Fð1; 106Þ¼139:48; p # 0:001). However, the young perceived a larger difference in odor intensity between the samples than the elderly (interaction of age group and aroma, Fð1; 106Þ¼13:84; p ¼ 0:007). The young tended to rate the odor as more intense than the elderly (Fð1; 106Þ¼3:67; p ¼ 0:058). No significant differ- ence was present in odor intensity ratings between the two tasting sessions. The ratings of flavor intensity were very similar to those of odor intensity. Both age groups rated flavor intensity as stronger in the heightened aroma sample (Fð1; 105Þ¼77:90; p # 0:001). The young perceived a larger difference in flavor intensity between samples than the elderly (interaction of age group and aroma, Fð1; 105Þ¼8:50, p ¼ 0:004). The flavor intensity ratings did not differ significantly between the first and second tasting sessions or between age groups. Hedonic ratings In the pleasantness ratings, the aroma concentration had a similar effect on both age groups in the first tasting session: the heightened aroma sample was rated as less pleasant than the regular aroma sample (Fð1; 106Þ¼54:90; p # 0:001). However, the difference between the pleasantness ratings of the two samples was smaller for the elderly than for the young (interaction of age group and aroma, Fð1; 106Þ¼ 19:55; p # 0:001). A change in the pleasantness ratings of the two age groups between the samples occurred from the first tasting session compared with the second one (Fig. 1). According to the young, the pleasantness ratings of the regular aroma sample were almost the same in both tasting sessions, while the pleasantness ratings of the heightened Fig. 1. Mean pleasantness ratings (þSEM) of the young ðn ¼ 58Þ and the elderly ðn ¼ 50Þ given to regular and heightened aroma samples in the tasting sessions. , regular aroma; B, heightened aroma. S. Koskinen et al. / Appetite 41 (2003) 87–9690 aroma sample decreased in the second tasting session compared with the first session. Among the elderly, the regular aroma sample was perceived to be slightly less pleasant in the second tasting session than in the first one, whereas the pleasantness of the heightened aroma sample increased in the second tasting session compared with the first one (interaction of age group and aroma and tasting, Fð1; 106Þ¼10:41; p ¼ 0:002). Home-use test Sensory ratings In home-use, the odor and flavor intensity ratings were affected by aroma concentration but not by age group. The odor intensity (Fð1; 106Þ¼13:95; p # 0:001) and flavor intensity (Fð1; 103Þ¼14:79; p # 0:001) of the heightened aroma sample were perceived as stronger in both age groups. Compared to the side-by-side tasting, both age groups rated less difference in odor and flavor intensity between the samples in the home-use (monadic testing). (Fig. 2a–d). Hedonic ratings The heightened aroma concentration decreased plea- santness ratings of both age groups also in home-use (Fð1; 105Þ¼21:48; p # 0:001). Again, the difference between the pleasantness ratings of the two samples was smaller for the elderly than for the young (interaction of age group and aroma, Fð1; 105Þ¼5:23; p ¼ 0:024). The pleasantness of the heightened aroma sample increased during the home-use, while the pleasantness of the regular aroma sample remained virtually unchanged (interaction of aroma and day, Fð2; 210Þ¼6:12; p ¼ 0:003) (Fig. 2e– f). Fig. 2. (a –f) Mean odor intensity ratings (þ SEM); (a– b), flavor intensity ratings (þSEM) (c–d), and pleasantness ratings (þ SEM) (e–f) of the young ðn ¼ 58Þ on the left (fig. a,c,e) and the elderly ðn ¼ 50Þ on the right (fig. b,d,f) given to the regular and heightened aroma samples in the tasting sessions and during home-use. , regular aroma; , heightened aroma. S. Koskinen et al. / Appetite 41 (2003) 87–96 91 Willingness and boredom ratings After the third and sixth days of home-use, subjects rated their willingness to eat the product again, to buy it and to recommend it to their friends. They also rated how bored they were with the product. Overall, the heightened aroma concentration decreased willingness to eat the product again (Fð1; 106Þ¼6:34; p ¼ 0:013), to buy it (Fð1; 106Þ¼8:12; p ¼ 0:005), and to recommend it to a friend (Fð1; 105Þ¼6:89; p ¼ 0:010). The heightened aroma concentration did not, however, affect willingness of the elderly consumers to eat the product again (interaction of aroma and age group, Fð1; 106Þ¼4:15; p ¼ 0:044), to buy the product (interaction of aroma and age group, Fð1; 106Þ¼3:89; p ¼ 0:051), or to recommend it (inter- action of aroma and age group, Fð1; 105Þ¼6:89; p ¼ 0:010). In general, the elderly were more willing to buy the product than the young (Fð1; 106Þ¼6:86; p ¼ 0:010). The boredom ratings were unaffected by either aroma concentration or age group. Intake Eighty-two percents of the young and 68 percents of the elderly ate at least 75% of the total amount served. There was no significant difference in the intake between the age groups, although the mean intake of the elderly was slightly lower. Both age groups ate significantly less of the heightened aroma than the regular aroma sample (Fð1; 93Þ¼13:04; p # 0:001). The intake of neither age group changed during the period of home-use. Olfactory test The young subjects detected (Fð1; 90Þ¼18:05; p # 0:001) and identified (Fð1; 90Þ¼53:12; p # 0:001) a larger number of odors than the elderly. The mean scores (SEM, range) for the young subjects were 15.8 (0.18, 14– 16) for detection and 13.5 (0.30, 9–16) for identification of a possible 16. The corresponding scores for the elderly were 14.7 (0.20, 8– 16) and 10.1 (0.35, 3 –15). Variation in the scores of the elderly was greater than in the scores of the young. In the elderly group, females scored higher in odor identification than males (Fð1; 37Þ¼6:10; p ¼ 0:018), their respective scores (SEM, range) being 10.7 (0.48, 3–15) and 8.2 (0.88, 5 –12). Gender had no significant effect on odor detection scores. Normative data are not yet available for ETOC, but the mean scores of the tested sample of elderly did not differ from the means of age-matched sub-groups in data on 1100 subjects (Thomas-Danguin et al., 2003b). Effect of olfactory performance and health status on hedonic ratings and intake of the elderly The detection and identification scores of the elderly were explored further, revealing an association between the scores and the number of diseases the elderly had. The mean detection score (SEM, range) for the elderly in the good health group (n ¼ 31; 23 females, 8 males, mean age 74.2 years) was 15.1 (0.31, 10–16) and for the poor health group (n ¼ 7; 6 females, 1 male, mean age 73.1 years) it was 13.0 (0.66, 8–16). The difference in the detection scores between the two groups was significant (Fð1; 36Þ¼8:31; p ¼ 0:007). The mean identification scores (SEM, range) were 10.5 (0.50, 5 –15) and 8.9 (1.04, 3 –13), respectively, but the difference in these scores was not significant. Hedonic ratings The pleasantness ratings of the ‘low score’ and ‘high score’ groups tended to differ from each other, but the difference was not significant (Fig. 3a). The elderly scoring low in odor identification rated pleasantness of the heightened aroma sample approximately at the same level during exposure, while their pleasantness ratings of the regular aroma sample decreased towards the end of home- use. Among the elderly scoring high in odor identification, the pleasantness ratings of the samples merged over exposure at home (Fig. 3b). Contrary to our expectations, ‘low score’ and ‘high score’ groups did not differ significantly in the rated odor intensity or flavor intensity of the samples. Fig. 3. (a–b) Mean pleasantness ratings (þSEM) (tasting sessions and home-use) of the elderly sub-grouped according to their scores in the odor identification (ETOC): low score group, n ¼ 19 (a); the high score group, n ¼ 20 (b). , regular aroma; , heightened aroma. S. Koskinen et al. / Appetite 41 (2003) 87–9692 Intake A significant difference was present in mean intake of both samples between the elderly scoring low (0– 10) and those scoring high (11– 16) in odor identification (Fð1; 31Þ¼4:55; p ¼ 0:041); the ‘low score’ group ate less than the ‘high score’ group (Fig. 4). There also seemed to be a connection between intake and number of nutritionally related diseases that the elderly subjects had. The subjects who had more diseases ate less of the snack product than their healthier counterparts. The mean intake of both samples in the elderly with good health tended to stay approximately at the same level, while the intake of the elderly with poor health tended to decrease during home-use (interaction of health status and day Fð2; 38Þ¼3:06; p ¼ 0:066) (Fig. 5). Discussion Four major findings emerged from this study: 1. During the exposures, the pleasantness ratings of the two samples merged because of the increasing popularity of the heightened aroma sample, especially in the elderly. 2. However, flavor enhancement did not increase intake among the elderly. 3. In general, the elderly were less responsive to differences in odor and flavor intensity of the samples. 4. The olfactory performance of the elderly was not associated with their hedonic responses. Sensory and hedonic responses to two flavor levels Although both age groups rated the odor and flavor intensity of the samples similarly during the testing period, their hedonic responses towards the samples differed. Both age groups initially preferred the regular aroma sample to the heightened aroma sample. In the young, the hedonic responses were relatively consistent, with the regular aroma sample being rated as more pleasant throughout the testing period. In the elderly, by contrast, the pleasantness ratings of the two samples merged and towards the end both samples were rated as equally pleasant. In a study by Philipsen, Clydesdale, Griffin, and Stern (1995), the hedonic responses of the elderly did not change with increasing aroma concentration, whereas the young responded strongly to a change in flavor intensity. Zandstra and de Graaf (1998) also found that variability in hedonic responses towards samples with different flavor intensity decreased with increasing age. Thus, the elderly may be more adaptive to slight changes in the flavor of food. Our hypothesis was that the elderly with poor olfaction would appreciate the sample with heightened aroma concentration more than the sample with regular aroma concentration. The hypothesis was based on studies by De Graaf et al. (1994, 1996) and Griep, Mets, and Massart (1997) and Schiffman and Warwick (1993), all of whom found that the elderly prefer food with added flavor, likely because it compensates for their diminished olfactory function. However, none of these studies actually measured the olfactory function of their subjects. Therefore, whether the preference for the stronger flavor was due to compen- sation for loss of smell remains unknown. In the present study, no connection between poor olfactory performance and appreciation of the heightened aroma concentration was found. The merging of pleasantness ratings given to the samples by the elderly suggests that during the test period, the elderly might have adapted to the intense flavor. Sensory losses occur gradually over a period of years, and the elderly may adjust their diets to the sensory loss (Rolls, 1993). Thus, with the snack products differing in flavor intensity, they first rated the regular aroma sample as more pleasant since the flavor intensity was more familiar to them, but with exposure they got used to the more intense flavor and thus slightly increased their liking for the initially less pleasant heightened aroma sample. The heightened aroma concentration of the sample decreased the willingness of the young subjects to eat Fig. 4. Mean intake (þSEM) (during home-use) of the elderly according to their scores in the odor identification (ETOC): low score group, n ¼ 16; high score group, n ¼ 17 , regular aroma; B, heightened aroma. Fig. 5. Mean intake (þSEM) (during home-use) of the elderly according to their health status: ¼ poor health group, n ¼ 7; ¼ good health group, n ¼ 31: S. Koskinen et al. / Appetite 41 (2003) 87–96 93 the product again, to buy it, and to recommend it to their friends, whereas the willingness ratings of the elderly were unaffected by the aroma concentration. This finding also indicates that the elderly are less sensitive to variations in flavor intensity than the young in their hedonic responses (Philipsen et al., 1995; Zandstra & de Graaf, 1998). Intake of the samples with two flavor levels In both age groups, the intake of the heightened aroma sample was lower than that of the regular aroma sample. The finding that the elderly also ate less of the heightened aroma sample is not congruent with earlier studies by Griep et al. (1997) and Schiffman and Warwick (1993), who found that flavor enhancement of food increased intake of the elderly. However, the evidence for the increase in intake of foods with enhanced flavor in these studies is not particularly compelling since the increase was not signifi- cant in most cases. This brings us back to the notion of Mattes and Cowart (1994) that increasing the appeal of foods by adding flavor is of little nutritional relevance. In the present study, the heightened aroma concentration may have been too high for both age groups, resulting in the intake being lower for the heightened aroma sample even among the elderly. According to Rolls (1992), an exces- sively strong flavor may decrease food intake. Vickers and Holton (1998) found that upon repeated consumption of ice tea subjects liked weaker tea better than strong tea, although initially they preferred the strong tea. During the exposure, they also consumed more of the weaker tea than the stronger one. In our study, the intake of neither of the two samples changed during home-use. According to Hetherington, Bell, and Rolls (2000), the way subjects regard the foods, i.e. whether they are considered to be staple or special foods, affects the development of monotony. In their study, the consumption of staple food (French fries) remained constant whereas the consumption of special food (chocolate) declined. Our subjects might have considered the snack product to be a staple food since it is yogurt-type product and thus the consumption was unaffected by repeated use. Olfactory performance as a predictor of intake and health status of the elderly The young were superior to the elderly in terms of odor detection and identification. This finding is in line with former studies (e.g. Hummel, Sekinger, Wolf, Pauli, & Kobal, 1997; Koskinen, Ka ¨ lvia ¨ inen, & Tuorila, 2003; Nordin, Nyroos, Maunuksela, Niskanen, & Tuorila, 2002). Furthermore, in the group of elderly, females outperformed males in the odor identification task. Similar results have been reported by Doty et al. (1984) and Nordin et al. (2002). We expected the elderly with low identification scores to appreciate the heightened aroma sample more than the regular aroma sample. After exposure, a slight trend occurred in that direction, however, the effect was not significant. The finding that indicates a failure of the increased flavor to increase pleasantness among the elderly with low odor identification ability may be due to the nature of the olfactory task that was conducted. The odor identification task requires memory and semantic proces- sing (Cain, de Wijk, Lulejian, Schiet, & See, 1998), and is not necessarily related to a person’s ability to perceive odor and flavor intensity, which affect the hedonic responses as well. However, current notions (Doty, Smith, McKeown, & Raj, 1994; Nordin et al., 2002) suggest that olfactory tests largely measure general olfactory performance, regardless of the type of the test. Furthermore, Tuorila and Koskinen (2003) found an indication, although a slight one, of predictive value of odor detection and identification tests in odor and flavor intensity ratings. The validity of different olfactory tests in the prediction of performance needs further attention. There was a tendency that the elderly in poor health (i.e. more reported diseases) performed more poorly in the olfactory test and ate less of the sample than the elderly in good health. Although the groups were not equal in size, the health status of the elderly did appear to have a role in sensory and nutritional issues. The poorer odor detection and identification ability may be due to the cognitive limitations of the subjects. The nutritionally related diseases may in part result from insufficient and unhealthy eating, which is more likely for subjects with poorer cognitive ability. According to Darnton-Hill (1992), physical, social, and psychological factors potentially contribute to nutri- tional problems of elderly people. However, these factors were not measured in this study. The trend of smaller intake of the subjects in poor health compared with those in good healths, although not significant, is of concern. Again, the total intake of the elderly with poorer performance in odor identification test was smaller than of those who performed better in the test. Thus, an association existed between intake and olfactory performance and health of the elderly. The subjects of this study were independently living and active elderly people. According to the literature, alongside aging, different diseases manifest that affect appetite, and the risk of developing anorexia of aging increases. However, anorexia among the aged is more common in institutionalized persons. Unexplained losses in food intake and body weight occurring near the end of life are characteristic of this syndrome (Horwitz, Blanton, & McDonald, 2002). Method Both the young and the elderly perceived greater differences in odor and flavor intensities in the tasting sessions than during home-use, probably because it is easier to find a difference in samples presented side-by-side, compared with monadic presentation, as in the home environment. According to La ¨ hteenma ¨ ki and Tuorila (1994), side-by-side presentation enables direct comparisons S. Koskinen et al. / Appetite 41 (2003) 87–9694 to be made of samples, thus emphasizing any differences between them. When a direct comparison is not allowed, respondents use their own expectations as a reference point (Cardello & Sawyer, 1992). Another reason for the larger difference between the samples may be that the evaluation conditions were more controlled in the tasting sessions than at home, enabling better concentration on the evaluations. The gender distribution of the subjects was unequal, since, due to availability, more females than males were included. However, gender had no significant effect on ratings of the samples. Hence, we may conclude that the uneven gender distribution did not affect results appreciably. The home-use period was conducted using a within- subject design, meaning that the subjects had to taste the same samples three times and rate the same attributes six times. A drawback was the possibility that the subjects would get bored over the course of the study, which would in turn affect their responses. According to Stubenitsky, Aaron, Catt, and Mela (1999), subjects may express their feeling of discontinued interest in the study itself instead of that of the products, when they are asked how bored they are with the studied products. In the present study, neither the boredom ratings increased nor the hedonic ratings decreased during exposure at home. It appears that the study period was not too long or monotonous to bore the subjects. Since this was a cross-sectional study, a generation effect might be present in the results. Behavioral and/or cultural factors may have influenced the responses of subjects. According to earlier studies (Pelchat & Schaefer, 2000; Tuorila, Andersson, Martikainen, & Salovaara, 1998), the elderly tend to rate samples more positively than the young. On the other hand, the elderly do express their dislike, when a sample does not appeal to them, as was the case in Koskinen et al. (2003). The usefulness of flavor enhancement of foods in changing food choices and nutrition behavior of the elderly needs further investigation, since there are ambiguous results and evidence for and against on that matter. Acknowledgements The authors thank practical trainees Linda van Mierlo and Pascalle Weijzen from Wageningen University, The Netherlands, for assistance with data collection. Bioferme Oy, especially the managing director Merja Scharlin, is thanked for providing samples and for co-operation. This study was financially supported by the European Commis- sion’s Quality of Life Fifth Framework Program QLK1-CT 1999-00010, project HealthSense: ‘Healthy eating: how changes in sensory physiology, sensory psychology, and socio-cognitive factors influence food choice’. References Cain, W., de Wijk, R., Lulejian, C., Schiet, F., & See, L.-C. (1998). Odor identification: Perceptual and semantic dimensions. Chemical Senses, 23, 309–326. Cardello, A. V., & Maller, O. (1982). Relationship between food preferences and food acceptance ratings. Journal of Food Science, 47, 1553– 1561. Cardello, A. V., & Sawyer, F. M. (1992). Effects of disconfirmed consumer expectations on food acceptability. Journal of Sensory Studies, 7, 253– 277. Darnton-Hill, I. (1992). Psychosocial aspects of nutrition and aging. Nutrition Reviews, 50(12), 476–479. De Graaf, C., Polet, P., & van Staveren, W. (1994). Sensory perception and pleasantness of food flavors in elderly subjects. Journal of Gerontology: Psychological Sciences, 49(3), 93–99. De Graaf, C., van Staveren, W., & Burema, J. (1996). Psychophysical and psychohedonic functions of four common food flavors in elderly subjects. Chemical Senses, 21, 293 –302. De Jong, N., De Graaf, C., & van Staveren, W. (1996). Effect of sucrose in breakfast items on pleasantness and food intake in the elderly. Physiology and Behavior, 60(6), 1453 –1462. Doty, R. L., Shaman, P., Applebaum, S. L., Giberson, R., Sikorson, L., & Rosenberg, L. (1984). Smell identification ability: Changes with age. Science, 226, 1441–1443. Doty, R. L., Smith, R., McKeown, D. A., & Raj, J. (1994). Tests of human olfactory function: Principal component analysis suggests that most measure a common source of variance. Perception and Psychophysics, 56(6), 701– 707. Drewnowski, A., & Shultz, J. M. (2001). Impact of aging on eating behaviors, food choices, nutrition, and health status. The Journal of Nutrition, Health and Aging, 5(2), 75–79. Griep, M. I., Mets, T. F., & Massart, D. L. (1997). Different effects of flavor amplification of nutrient dense foods on preference and consumption in young and elderly subjects. Food Quality and Preference, 8(2), 151– 156. Griep, M. I., Mets, T. F., & Massart, D. L. (2000). Effects of flavor amplification of Quorn w and yoghurt on food preference and consumption in relation to age, BMI and odour perception. British Journal of Nutrition, 83, 105–113. Hetherington, M. M., Bell, A., & Rolls, B. J. (2000). Effects of repeated consumption on pleasantness, preference and intake. British Food Journal, 102(7), 507– 521. Hoffman, H. J., Ishii, E. K., & Macturk, R. H. (1998). Age-related changes in the prevalence of smell/taste problems among the United States adult population—Results of the 1994 disability supplement to the National Health Interview Survey (NHIS). Olfaction and taste XII. Annals of the New York Academy of Sciences, 855, 716–722. Horwitz, B. A., Blanton, C. A., & McDonald, R. B. (2002). Physiologic determinants of the anorexia of aging: Insight from animal studies. Annual Reviews of Nutrition, 22, 417– 438. Hummel, T., Sekinger, B., Wolf, S. R., Pauli, E., & Kobal, G. (1997). Sniffin Sticks: Olfactory performance assessed by the combined testing of odor identification, odor discrimination and olfactory threshold. Chemical Senses, 22, 39 –52. Koskinen, S., Ka ¨ lvia ¨ inen, N., & Tuorila, H. (2003). Perception of chemosensory stimuli and related responses to flavored yogurts in the young and elderly. Food Quality and Preference, in press. La ¨ hteenma ¨ ki, L., & Tuorila, H. (1994). Liking for ice cream measured with three procedures: Side-by-side, after consumption and single samples. Journal of Sensory Studies, 9, 455– 465. Mattes, R. D., & Cowart, B. J. (1994). Dietary assessment of patients with chemosensory disorders. Journal of the American Dietetic Association, 94(1), 50– 56. Murphy, C. (1993). Nutrition and chemosensory perception in the elderly. Critical Reviews in Food Science and Nutrition, 33(1), 3–15. S. Koskinen et al. / Appetite 41 (2003) 87–96 95 Nordin, S., Nyroos, M., Maunuksela, E., Niskanen, T., & Tuorila, H. (2002). Applicability of the Scandinavian Odor Identification Test: A Finnish– Swedish comparison. Acta Otolaryngologica, 122, 294–297. Pelchat, M. L., & Schaefer, S. (2000). Dietary monotony and food cravings in young and elderly adults. Physiology and Behavior, 68, 353 –359. Philipsen, D. H., Clydesdale, F. M., Griffin, R. W., & Stern, P. (1995). Consumer age affects response to sensory characteristics of a cherry flavored beverage. Journal of Food Science, 60(2), 364–368. Porcherot, C., & Issanchou, S. (1998). Dynamics of liking for flavored crackers: Test of predictive value of a boredom test. Food Quality and Preference, 9, 21–29. Rolls, B. J. (1992). Aging and appetite. Nutrition Reviews, 50(12), 422– 426. Rolls, B. J. (1993). Appetite, hunger, and satiety in the elderly. Critical Reviews in Food Science and Nutrition, 33(1), 39–44. Rowe, J. W., & Kahn, R. L. (1987). Human aging: Usual and successful. Science, 237, 143–149. Schiffman, S. S. (1994). Changes in taste and smell: Drug interactions and food preferences. Nutrition Reviews, 52(8), S11 –S14. Schiffman, S. S. (1997). Taste and smell losses in normal aging and disease. Journal of American Medical Association, 278(16), 1357 –1362. Schiffman, S. S., & Warwick, Z. S. (1993). Effect of flavor enhancement of foods for the elderly on nutritional status: Food intake, biochemical indices, and anthropometric measures. Physiology and Behaviour, 53, 395– 402. Simons, L. A., McCallum, J., Friedlander, Y., & Simons, J. (2000). Healthy ageing is associated with reduced and delayed disability. Age and Ageing, 29, 143–148. Stubenitsky, K., Aaron, J. I., Catt, S. L., & Mela, D. J. (1999). Effect of information and extended use on the acceptance of reduced-fat products. Food Quality and Preference, 10, 367– 376. Thomas-Danguin, T., Rouby, C., Sicard, G., Vigoroux, M., Barkat, S., Brun, V., Farget, V., Johansson, A., Bengtzon, A., Hall, G., Rousseau, F., Dumont, J-P., Ormel, W., Essed, N., de Graaf, C., Gourillon, S., Issanchou, S., Hover, S., Zunft, F., Nielsen, R., Koskinen, S., & Tuorila, H. (2003). ETOC: A short olfactory test culturally validated across Europe. Chemical Senses, 28(1), E60. Thomas-Danguin, T., Rouby, C., Sicard, G., Vigoroux, M., Farget, V., Johansson, A., Bengtzon, A., Hall, G., Ormel, W., de Graaf, C., Rousseau, F., & Dumont, J.-P. (2003). Development of ETOC: A European test of olfactory capabilities. Rhinology,in press. Tuorila, H., & Koskinen, S. (2003). Olfactory identification and detection tests in prediction of odor and flavor intensity ratings in the elderly and young. Chemical Senses, 28(1), E62. Tuorila, H., Andersson, A ˚ ., Martikainen, A., & Salovaara, H. (1998). Effect of product formula, information and consumer characteristics on the acceptance of a new snack food. Food Quality and Preference, 9(5), 313– 320. Tuorila, H., Niskanen, T., & Maunuksela, E. (2001). Perception and pleasantness of food with varying odor and flavor among the elderly and young. The Journal of Nutrition, Health and Aging, 5(4), 266 –268. Vickers, Z., & Holton, E. (1998). A comparison of taste test ratings, repeated consumption, and postconsumption ratings of different strengths of iced tea. Journal of Sensory Studies, 13, 199–212. Wysocki, C. J., & Pelchat, M. L. (1993). The effects of aging on the human sense of smell and its relationship to food choice. Critical Reviews in Food Science and Nutrition, 33(1), 63–82. Zandstra, E. H., & de Graaf, C. (1998). Sensory perception and pleasantness of orange beverages from childhood to old age. Food Quality and Preference, 9(1/2), 5–12. S. Koskinen et al. / Appetite 41 (2003) 87–9696 . No diseases ðnÞ Young 58 3/ 55 18– 34 23. 1 2 0 0 56 Elderly 50 12 /38 63 85 73. 7 20 13 7 20 S. Koskinen et al. / Appetite 41 (20 03) 87–96 89 presented. particular food 0195-66 63/ 03/ $ - see front matter q 20 03 Elsevier Ltd. All rights reserved. doi:10.1016/S0195-66 63( 03) 00054-0 Appetite 41 (20 03) 87–96 www.elsevier.com/locate/appet