Presented at the 20th Meeting of the Association for Chemoreception Sciences, Sarasota, Florida, April 25, 1998.
Objective To develop a simple test for the screening of gustatory function in clinical settings.
Study Design We tested 101 healthy volunteers (44 male and 57 female volunteers; mean age, 47 y) with the following gustatory test: the substances sucrose (sweet), citric acid (sour), sodium chloride (salty), and caffeine (bitter) were presented as tablets (diameter 4 mm) similar to common sweetener tablets. For quantitative assessment of whole-mouth gustatory function we used six different dosages with dilutions of each tastant in 50% steps. The highest dosage could be easily detected (sucrose, 30 mg; citric acid, 3 mg; sodium chloride, 2 mg; caffeine, 2 mg), and the lowest concentration was within threshold range.
Methods Twenty-eight tablets (si- different dosages of the four basic tastes plus four tasteless tablets) were tried in a randomized order. The entire test required 15 to 20 minutes. To evaluate the within-subject test-retest reliability, sessions were repeated after 1 week. Results were compared with those obtained by means of a conventional three-drop, forced-choice procedure using the method of ascending limits.
Results Results of the new gustatory test were significantly correlated with those obtained using the three-drop, forced-choice procedure (correlation coefficient [r] = 0.66, P < .001). In general, women performed better than men. Furthermore, younger subjects exhibited a significantly higher gustatory sensitivity in both tests compared with older subjects.
Conclusions This quantitative test of whole-mouth gustatory function is easy to use, can be self-administered, requires little time, and has a long shelf-life. It appears to be suited for routine clinical assessment of gustatory function.
Perception of taste can be altered by various disorders or therapeutic regimens. For example, a number of diseases, such as upper respiratory tract infections, colds, or dental procedures, are associated with taste dysfunction. 1 Head trauma patients frequently complain about alterations in their ability to taste. 2 Drug-related gustatory disorders may occur during treatment with such drugs as phenylbutazone, carbamazepine, baclofen, levodopa, dapsone, or terbinafine. 3 In additional, radiotherapy and chemotherapy (e.g., vinblastin 4,5) may produce changes in taste perception.
Psychophysical tests used for the assessment of changes in gustatory sensitivity follow different approaches. They can roughly be divided into regional tests of gustatory function and whole-mouth tests. Regional testing is frequently applied in specialized centers 6 to evaluate taste function in specific areas of the tongue. This technique allows the detailed examination of innervation fields of the tongue by the cranial nerves. Traditionally, liquids are administered to the tongue. 2,7 However, the test solution spreads all over the tongue and is diluted immediately after application of a drop, which may be a problem for quantitative evaluation. This method was improved by the use of filter-paper disks. 8 They are soaked with a liquid and allow application of the tastant without diffuse spreading or immediate dilution of the tastant. Another method is based on the use of cotton-tipped swabs. The swabs are used to “paint” taste solutions onto different areas of the tongue. 9 However, in all these studies taste solutions have to be freshly prepared. With regard to shelf life, taste tests based on edible wafers 10 appeared to be a major breakthrough.
Regional testing has clear advantages in clinical settings. It allows the detailed assessment of localized deficits 11 resulting from lesions of the chorda tympani, which may be due to surgery, tumors, or localized viral infections. However, whole-mouth tests of gustatory function are needed to assess “everyday” taste experiences that are not reflected by regional tests. In fact, a number of tests have been described that allow assessment of whole-mouth gustatory function. However, these tests have the following major drawbacks. 1) Similar to regional tests of gustatory function, whole-mouth tests use liquids that are brought onto the patient's tongue. 12 However, the use of liquids has inherent difficulties such as the microbial contamination of the aqueous solutions. 2) Most taste-testing protocols require at least 1 hour of patient-staff contact. 2 3) Most of these tests have not been investigated with regard to their test-retest reliability.
Because these factors limit the routine clinical application of established tests, the aim of the study was to develop a highly practical method for gustatory testing. Our idea was to use tablets tasting sweet, sour, salty, and bitter. To allow quantitative gustatory testing, different concentrations of the tastants were presented in a pseudo-randomized order. A newly developed protocol limited the time required for testing to 15 to 20 minutes.
The technique's validity should be evaluated with regard to its test-retest reliability and the correlation of results with an established taste test, the “three-drop method.” We investigated sensory differences related to age and sex as additional criteria of the test's validity. Based on previous investigations, 10,13,14 it was hypothesized that women and younger subjects exhibit a higher gustatory sensitivity than men and older subjects.
MATERIAL AND METHODS
Preparation of Taste Tablets
Presentation of the four basic tastes was performed by means of tablets similar to common sweetener tablets. The round tablets were white, with a diameter of 4 mm. To produce sweet, salty, sour, and bitter taste, sucrose, sodium chloride, citric acid, and caffeine, respectively, were added to the tablets. For quantitative assessment of gustatory function, six different dosages (in 50% steps) of each substance were used (Table I). The dosages of the tablets were determined during preliminary experiments in healthy subjects recruited from the laboratory staff. In this pilot study the highest dosage of the four tastants was detected by all subjects; in contrast, the lowest concentration was detected in only a few instances.
Table Table 1.. Dosages in Milligrams of the Tasting Substances Administered in the Tablets.
Microcrystalline cellulose (Avicel PH101, Lehmann & Voss & Co., Hamburg, Germany), polyethyleneglycol 6000 (Lutrol E 6000, BASF, Ludwigshafen, Germany), and Aerosil 200 (Degussa, Frankfurt, Germany) were used as additional ingredients, which are similar to what is found in registered pharmaceutical products. The tablets were produced on an Excenter press (Fette Exakta 1, Wilhem Fette GmBH, Schwarzenbek, Germany) using a punch with 4-mm diameter in the laboratory of the Institute of Pharmaceutical Technology (Prof. Dr. G. Lee) of the University of Erlangen-Nürnberg (Erlangen, Germany).
The correct dosages were monitored by means of different techniques (sucrose, enzymatic analysis, Boehringer Mannheim [Penzberg, Germany] test kit; citric acid, enzymatic analysis, Boehringer Mannheim test kit; sodium chloride, ion-selective electrode and flame photometry; caffeine, high-performance liquid chromatography). With respect to the concentrations of the tastants in the tablets, current regulations of the German Pharmacopoeia (DAB 10) were applied. According to these regulations, a variation of ±15% of the dosages was regarded to be acceptable for use in the gustatory test.
Twenty-eight tablets (six different dosages of the four tastants plus four tasteless tablets) were stored in white boxes labeled with numbers from 1 to 28 (Fig. 1). Tablets were placed on the subject's tongue in pseudo-randomized orders using a pair of tweezers; the order of application was randomized across subjects.
Subjects were instructed to chew on the tablets as long as necessary to identify the taste. They rinsed with low-sodium, noncarbonated mineral water. The interval between presentation of the tablets was approximately 30 seconds. For each tablet, subjects were asked to choose among one of five descriptors, namely, “sweet, sour, salty, bitter, no taste.” Completion of the test required approximately 15 to 20 minutes.
One hundred one volunteers (44 men, 57 women; ages, 21–81 y; mean age, 47 years) participated in the study. All subjects were in excellent health. None of them reported olfactory or gustatory dysfunction. They were instructed to refrain from smoking, drinking coffee, eating, and chewing gum for at least 1 hour before the experiment. The protocol was explained in detail to the subjects, and written consent was obtained. The study was performed in accordance to the Declaration of Helsinki/Hong Kong. To compare the results of the new test method with an established method, subjects had to perform the following two gustatory tests.
Three-drop, forced-choice procedure.
A series of three drops was placed in the middle of the subject's extended tongue. One of the drops contained the tastant; the other two drops were water. Taste solutions were prepared in eight dilutions in 50% steps starting from sucrose (300 mg/mL for sucrose, 60 mg/mL for citric acid, 80 mg/mL for sodium chloride, and 20 mg/mL for caffeine). Administration in ascending concentrations started with the highest dilution. Concentrations were increased until correct recognition of the taste drop occurred two times in a row. This concentration was determined as threshold. The entire procedure for the four tastants took approximately 25 minutes.
Following the drop method, the new tablet method was applied. Subjects had to identify the taste of 24 (4 × 6) tablets of the four tastants and four (+ 4) tasteless tablets (= 28 tablets), which were administered in a pseudo-randomized order, as described above.
To evaluate the within-subject test-retest reliability, sessions were repeated after 1 week. Results of the new tablet method were compared with those obtained by means of the three-drop, forced-choice procedure.
Results were analyzed using SPSS 6. Data were submitted to a multivariate ANOVA (MANOVA, repeated measures design) using within-subject factors “taste test” and “session,” and between-subject factors “age group,” “sex,” and “smoking.” Estimates of significance were based on the averaged F test; degrees of freedom (df) were adjusted according to Greenhouse-Geisser. In case the MANOVA yielded significant results (α level = 0.05), Bonferroni adjusted Student t tests were performed. In addition to the F value, df, and the level of significance (P), we also present the observed power at 0.05 level (power). Coefficients of correlations (Spearman) were computed for evaluation of test-retest reliability and correlations of the results of the two methods.
Percentage of Correct Ratings in Tablet Method
“Sweet” was identified correctly by nearly all of the 101 participating subjects in dilution steps 1 and 2 (Fig. 2). Dilution steps 5 and 6 were correctly identified by only 29.7 (8.9%) of the subjects, respectively. At lower concentrations (dilution steps 4 and 5), it was mostly mistaken for sour.
“Sour” was also identified by most of the subjects when highest concentrations had been presented. In dosages around threshold, it was mostly confused with “salty” (13.4% of the subjects confused the two in dilution step 4; 16.3% of the subjects did so in dilution step 5).
“Salty” was more difficult for the subjects to identify. At lower dosages, confusion occurred frequently with “sour” (15.8% of subjects in dilution step 4; 18.3% in dilution step 5) and “bitter” (11.9% of subjects in dilution step 4; 19.3% in dilution step 5).
“Bitter” was identified correctly by most of the subjects in the higher concentrations, and performance decreased with higher dilutions. Confusion occurred with “salty” (6.9% of subjects in dilution step 4; 8.9% in dilution step 5), but also with “sour” and “sweet.”
To evaluate the test-retest reliability with regard to performance in the first and second sessions, scores were determined as the percentage of correct identification of the dilution steps of the four taste qualities. In case all six dilution steps of the four taste qualities were identified correctly (total of 24 items), it was noted as a score of 100%. The results of sessions 1 and 2 exhibited a significant correlation (r = 0.69, P < .001).
The three-drop method was employed as a reference test to the new tablets method. Performance in the three-drop method was expressed in a composite score determined as the mean threshold (expressed as dilution steps) obtained for the four taste qualities. Results of sessions 1 and 2 exhibited a significant correlation (r = 0.71, P < .001). The factor “session” had no significant effect (F = 0.38;df = 1;P = .541; power = 0.056), indicating that there was no difference in performance between sessions 1 and 2.
Correlation between tablet and drop methods.
No statistical difference was observed regarding the subjects' performance in the two tests (F = 0.20;df = 1;P = .653; power = 0.048). The significant correlation between the two tests (r = 0.66, P < .001) indicated that the taste tablets were suited to assess interindividual changes in gustatory sensitivity.
Effects of Age
When comparing the results of the younger and older subjects (ages 21–41 y and 42–81 y, respectively), a significant effect of the factor “age” could be observed for both tablets and drops (F = 29.18;df = 1;P < .001; power = 1.00) (Fig. 3). For the tablets, consequently performed Student t tests revealed that performance decreased in the older age group for “sour” and “salty” (t >3.54;df = 99;P < .01) (Fig. 3 A). For the drops, a significant difference between the two age groups could be observed for “sour,” “salty,” and “bitter” (t > 3.331;df = 99;P < .01) (Fig. 3 B). In general, scores obtained for all four qualities decreased as a function of age (Fig. 4), regardless of the test used.
Effects of Sex
Women were generally found to be better than men in terms of gustatory sensitivity in both test methods (F = 5.59;df = 1;P < .05; power = 0.645). However, when applying post hoc statistics, this finding only became significant in the three-drop method for the taste quality “bitter” (t = 2.64;df = 99:P < .001), not for the other taste qualities.
Difference Between Smokers and Nonsmokers
No significant differences between smokers and nonsmokers could be detected.
The present study investigated a new test for the screening of gustatory function. By testing the subjects on two occasions, we obtained data indicating the reliability of the test. The test-retest correlation of 0.69 is comparable to other gustatory tests 15,16; it clearly indicates a good reproducibility of the results. In addition, the validity of the new test could be shown by the significant correlation between results for the tablets and drops.
Mean thresholds obtained with the drop method are in close agreement with published data. 17–19 However, thresholds measured with the tablets were generally higher than those obtained with the drops. An explanation for these results may be the way of presentation of the tastant. The drop method allows delivery of the tastant already dissolved in water. In contrast, the tablet method requires saliva or water for the tastant to dissolve. Thus the concentration of the dissolved tastant may be lower compared with administration of drops. Therefore higher dosages of the tastants might be required to be detected as a tastant.
The age-related decrease of gustatory function is not only of importance in terms of the enjoyment of food; also, taste disorders may lead to inadequate food intake and malnutrition. The present data indicate that gustatory function is decreased in older subjects. These results are similar to the findings of Schiffman 20 and Hummel et al. 10 Evidence appears to accumulate indicating that both smell and taste exhibit a life-span decrease. However, the gustatory system seems to be more robust compared with the sense of smell. 19 There, functional anosmia is frequently seen, while ageusia is relatively rare. The observed lack of distinct age-dependent changes in taste function may be the result of methodological problems (i.e., a small number of subjects used in some of the studies or the relative insensitivity of the applied gustatory test).
It has been reported that age-related decreases in taste perception can be diminished after control for saliva production. 14 Because saliva production was not controlled in this experiment, it seems possible that the observed age-related change may be the result of other age-related processes. However, what argues against this view is that none of the subjects complained of dry mouth or insufficient saliva production. In addition, recent experiments in patients with xerostomia indicated that saliva replacement therapy did not improve gustatory function. 21
With regard to sex, our results show that women generally exhibit a more sensitive sense of taste than male subjects. This difference between the sexes is similar to previous findings. 10,13,14 Similarly to the sense of smell, women appear to be more sensitive to chemical stimuli than males.
Reasons for age-related sex differences in taste perception are not entirely clear. 22 When searching for explanations, the first idea is related to the differences in the endocrine system between males and females that could indirectly influence the gustatory system. It was shown that gonadectomy in female rats could eliminate taste aversion to quinine sulfate; administration of the hormones estrogen and progesterone abolished this effect. 23 Therefore differences in hormonal status may be responsible for the observed differences in taste perception. However, changes in taste perception during the menstrual cycle could be demonstrated in only a few instances. 24 Concerning the perception of different taste qualities, our findings support previous investigations that women had a better perception of “bitter” stimuli than men. This increased sensitivity may serve as an explanation for why it was reported that women dislike bitter foods more than men. 22 Besides the hormonal differences, there may be a number of other reasons why sex-related differences have been reported earlier. 22 They could result from a sampling bias (nonrandom selection of subjects) or variable taste receptor sites. It was also reported that men have larger salivary glands than women, which may lead to altered taste perception. 25
Interestingly, we noticed a frequent confusion of salty and sour tastes. This may be due to the similarity of the two taste qualities, because they both sting slightly on the tongue. It is well known that sour and salty tastes are associated with a certain degree of irritation. 26 Furthermore, both tastants are frequently used in combination in foods with rich flavor (e.g., in a salad sauce [sour, vinegar; salty, common salt]). This well-known combination and the similarity of the two tastants with regard to irritation of the tongue could be a reason for the mutual confusion of these two taste qualities.
In contrast, sweet taste seems to be devoid of irritation. 27 Therefore it was not confused so easily with other taste qualities and could be better distinguished from the other tastants.
Furthermore, there are reports that a considerable proportion of individuals in the general population exhibit “sour-bitter” confusions. 28 We could not generally support these findings in our investigations. Even though “sour-bitter” confusions occurred, it appeared that “bitter” was more often confused with “salty,” but also with “sweet.”
The present data in healthy subjects suggest that gustatory function can be evaluated with this new test. Advantages of the test are its portability, long shelf-life, and quick administration.
We thank Prof. Dr. G. Lee (Institute of Pharmaceutical Technology) for his generous support of the study. Special thanks also goes to Mr. Distler of the Institute of Legal Medicine for analyzing concentrations of caffeine in the tablets.