Self‐Administered Taste Testing Without Water: Normative Data for the 53‐Item Waterless Empirical Taste Test (WETT)

ABSTRACT

Background

According to the National Health and Nutrition Examination Survey (NHANES), more than 20 million Americans suffer from chronic taste problems. Unfortunately, quantitative testing is rarely performed, self‐reports are inaccurate, and “taste” is often confused with olfaction‐dependent flavors. Clearly, a reliable, practical, and norm‐based clinical taste test is sorely needed.

Methods

The recently developed 53‐item Waterless Empirical Taste Test (WETT) was self‐administered by 1392 healthy subjects (493 men and 899 women) spanning the ages of 10–94. Normative percentile ranks were developed, and the influences of sex, age, COVID‐19 infection history, and smoking behavior were evaluated using general linear models.

Results

On average, WETT scores declined with age beginning in the 30s and were lower in men than in women (ps < 0.0001), with the sex difference most apparent between 40 and 70 years of age. Subtest scores for the five basic taste qualities followed similar age‐ and sex‐related patterns. No influences of COVID‐19 history were evident. Smokers had lower salty (NaCl) and bitter (caffeine) WETT scores than nonsmokers. The developed normative tables make it possible to determine a patient's absolute dysfunction score as well as a relative score based on age‐ and sex‐related percentiles.

Conclusion

This study provides normative data based on 1392 healthy subjects for interpreting a patient's performance on a reliable self‐administered taste test that requires no liquid stimuli or rinses. Its findings support the view that both sex and age are significant determinants of taste function in the general population.

1. Introduction

The sense of taste largely determines whether a food or beverage should be ingested and plays an important role in quality of life, esthetics, nutrition, and safety. According to data from the National Health and Nutrition Examination Survey (NHANES), more than 20 million Americans suffer from chronic taste problems [1]. This primary sensory system is negatively impacted by a range of diseases and disorders, including cystic fibrosis [2], traumatic brain injury [3], head and neck radiotherapy [4], chronic rhinosinusitis [5], early stage cancers [6], hypertension [7], hypothyroidism [8], diabetes [9], obesity [10, 11, 12], kidney disease [13], Parkinson's disease [14], multiple sclerosis [15], iatrogenesis (e.g., bariatric surgery) [16], and liver disease [17], to name a few.

Among the most common causes of taste disturbances are such widely prescribed medications as antifungal agents [18] and cardiovascular ACE inhibitors and beta‐blockers [19]. Recent data suggest that GLP‐1 agonists used for weight loss depress taste test scores [20]. Unfortunately, without testing, most patients are unaware of their gustatory dysfunction [21]. Interestingly, in cases of long‐term damage to taste nerves from third molar extractions, there is objective evidence that hypersensitivity, rather than hyposensitivity, occurs [22].

Unlike the situation for olfaction where sensitive, practical, and validated self‐administered quantitative tests are widely employed [23, 24, 25, 26, 27, 28], similar types of taste tests are not generally available. An exception is the recently developed 53‐item Waterless Empirical Taste Test (WETT) [29]. This reliable test (test‐retest r = 0.92) makes it possible to measure taste function without the intervention of a test administrator. All five major taste qualities are evaluated, and the test concentrations are designed psychometrically to provide an overall normal distribution of test scores. Since it requires no stimulus preparation and employs neither liquid tastants nor liquid rinses, it can be used in a range of non‐laboratory conditions, including such non‐conventional environments as submarines and space stations. Its ability to be self‐administered and sent through the mail has made it very convenient for general population studies, as well as ones involving COVID‐19 disease patients [30, 31].

This paper provides age‐ and sex‐related normative data derived from the self‐administration of this test by a large number of healthy participants. These norms make it possible, for the first time, to rapidly and accurately determine whether a patient has a true taste deficit, both in absolute terms relative to younger persons, and in terms relative to a patient's sex‐ and age‐related peers. Importantly, self‐administration is efficient and cost effective, eliminates the need for direct physician or technician involvement, and precludes potential examiner influences on test results.

2. Materials and Methods

2.1. Participants

The study population was comprised of 1392 healthy participants 10–94 years of age (899 women, 493 men; respective mean ages [95% CI] = 38.63 [37.61, 39.65] and 37.88 [36.49, 39.27]). Of these, 1030 never smoked cigarettes (74%), 229 were past smokers (16%), and 133 were current smokers (10%). Most were recruited from advertisements placed on a social media website (Reddit) or local bulletin boards, and the test materials were sent to them by mail and returned to the investigators after completion. Each participant was paid $30. The ads sought healthy persons who can “…help in creating normative data for smell and taste tests which will be used to track quality control and assess test‐retest reliability.” Individuals with a self‐reported history of a disorder known to impact smell or taste function (e.g., head trauma and neurodegenerative disease) were excluded. Those with a past history of COVID‐19 (n = 673) were not excluded in light of evidence that WETT scores are not impacted by such a history [31]. Participants came from 49 US states except for Hawaii. Four participants came from the Virgin Islands and one from Puerto Rico. All participants provided informed written consent, and the study was approved by the WIRB‐Copernicus Group Independent Review Board. This research complies with the Declaration of Helsinki for medical research involving human participants.

2.2. Test Details and Administration Procedures

The WETT was self‐administered by the participants according to written instructions [33]. Each of the 53 test items is comprised of a 1 × 6 cm plastic taste strip. A 1 × 2.5 cm monomer cellulose pad is present on one side of each strip, containing either dried sucrose (0.20, 0.10, 0.05, or 0.025 g/mL), citric acid (0.025, 0.05, 0.10, or 0.20 g/mL), sodium chloride (NaCl; 0.0313, 0.0625, 0.125, or 0.25 g/mL), caffeine (0.011, 0.022, 0.044, or 0.088 g/mL), monosodium glutamate (MSG; 0.017, 0.034, 0.068, or 0.135 g/mL), or no stimulus. As shown in Figure 1, the strips are provided in four convenient packs from which each strip is removed in a consecutively numbered sequence.

FIGURE 1.

Picture of stimulus response sheets (left) and four packs of self‐administered WETT taste strips (right). Each strip is pulled from a friction holding strip at the bottom of the set of strips. See text for details. Courtesy of Sensonics International. Copyright 2021 Sensonics International.

On a given trial, each subject was instructed to place the appropriate cellulose pad on the tongue and to move it over the tongue's surface, including its tip and edges for 5–10 s. The participant then marked his or her response on the designated line of the response form, indicating which taste sensation was perceived (sweet, sour, salty, bitter, and brothy) or whether no taste was present (Figure 1).

In the WETT test sequence, the four concentrations of each stimulus are presented twice (eight total trials/tastant). In the first half of the test, the concentrations ascend from weak to strong with the different tastants presented in a random order with no tastant immediately following itself. In the second half of the test, the reverse presentation order is made, that is, going from strong to weak concentrations.

The WETT is a general measure of the ability to identify various concentrations of the five basic taste qualities with a minimum number of trials and without the calculation of thresholds, per se. Its presentation paradigm, however, is operationally similar to that of a recognition threshold test given its sequential ascending and subsequent descending stimulus concentration presentations. Unlike traditional recognition thresholds, however, the concentrations of a given tastant do not immediately follow one another, but are interspersed among those of the other stimulants, resulting in efficient stimulus presentations with minimal likelihood of adaptation. This approach to taste testing has been employed by others [34, 35, 36].

2.3. Statistical Analyses

General linear models (GLMs) were employed to determine the influences of age, sex, COVID‐19 history (yes, no), and current smoking behavior (yes, no) on the total WETT score and on each of the subtest scores, that is, sweet (sucrose), sour (citric acid), bitter (caffeine), salty (NaCl), and brothy (MSG) [37]. Data were entered into the models using a backward stepwise algorithm with alpha‐to‐enter and alpha‐to‐remove values set at 0.15. Subsequently, separate norms were developed for total test scores and for the subtests, with percentiles calculated on the scores for each sex and seven age categories.

3. Results

3.1. Variables Affecting Taste Test Scores

Only age and sex contributed significantly to the total 53‐item WETT score (age F [1,1389] = 77.68, p < 0.0001, η ² = 0.051; sex F [1, 1389] = 37.04, p < 0.0001, η ² = 0.026). Scores decreased with age (Figure 2) and women exhibited higher average scores than men (respective means [95% CIs] = 35.16 [34.56, 35.76] and 32.16 [31.35, 32.97])—an age‐related disparity that was largest for those between the ages of 40 and 69 years (Table 1). Non‐smokers tended to outperform current smokers, but the effect was not significant at the 0.05 level (p = 0.12).

FIGURE 2.

Mean (95% CI) age‐related changes in WETT scores based upon 1392 healthy subjects (493 men and 899 women). See text for details.

TABLE 1.

Mean (95% CI) total WETT scores for seven age groups for men and woman.

	Age group
	10–19	20–29	30–39	40–49	50–59	60–69	70+
Men	34.04 (31.86, 36.23)	35.59 (34.04, 37.14)	33.45 (31.87, 35.03)	29.72 (27.93, 31.51)	27.92 (25.05, 30.78)	28.09 (25.27, 30.91)	28.44 (24.49, 32.38)
Women	36.08 (33.97, 38.19	37.72 (36.60, 38.85)	35.52 (34.17, 36.88)	34.34 (33.07, 35.60)	33.63 (31.87, 35.40)	31.14 (30.22, 34.06)	30.03 (27.12, 32.93)
Mean difference	2.84	2.13	2.07	4.62	5.71	3.05	1.59
p value	0.207	0.027	0.054	0.001	0.001	0.017	0.502
Male sample size	46	124	119	100	48	33	23
Female sample size	76	225	186	209	101	65	37
Total sample size	122	349	305	309	159	98	60

Note: p values are based on one‐way analyses of variance. Significant p values are in bold. See text for details.

The mean (95% CI) data for each of the five WETT subtests as a function of sex and age are presented in Table 2. GLMs found the average scores to be significantly higher in women than in men for all taste qualities (ps ≤ 0.0001) except for brothy (MSG)(p = 0.176). As is apparent from this table, the differences were more marked in some age groups than others. Both salty (NaCl) and bitter (caffeine) WETT scores were significantly lower in cigarette smokers than in non‐smokers (salty means [95% CIs] = 5.11 [4.72, 5.49] and 5.56 [5.44, 5.69], F [1,1388] = 5.73, p = 0.017, η ² = 0.004; bitter means [95% CIs] = 4.12 [3.65, 4.59] and 4.66 [4.51, 4.81], F [1,388] = 4.42, p = 0.036].

TABLE 2.

Mean (95% CI) subtest WETT scores for seven age groups for men and woman.

Men: Age group
	10–19	20–29	30–39	40–49	50–59	60–69	70+
Sweet	4.97 (4.33, 5.62)	5.11 (4.74, 5.49)	4.36 (3.96, 4.77)	3.98 (3.45, 4.51)	3.41 (2.59, 4.25)	3.58 (2.62, 4.53)	3.44 (2.24, 4.63)
Sour	5.83 (5.39, 6.26)	5.92 (5.56, 6.28)	5.58 (5.24, 5.92)	4.73 (4.30, 5.16)	4.46 (3.79, 5.12)	4.06 (3.29, 4.83)	4.04 (3.00, 5.09)
Bitter	4.13 (3.37, 4.90)	4.91 (4.44, 5.38)	4.38 (3.92, 4.84)	3.32 (2.81, 3.83)	4.04 (3.31, 4.76)	2.85 (1.90, 3,80)	2.26 (1.28, 3.25)
Salty	5.52 (4.93, 6.12)	6.03 (5.73, 6.34)	5.53 (5.16, 5.90)	4.84 (4.42, 5.26)	5.19 (4.58, 5.79)	4.24 (3.44, 5.05)	5.22 (4.52, 5.92)
Umami	2.76 (2.07, 3.45)	3.30 (2.85, 3.74)	3.12 (2.65, 3.58)	2.32 (1.84, 2.80)	2.48 (1.86, 3.10)	2.24 (1.50, 2.98)	2.30 (1.20, 3.41)
Sample size	46	124	119	100	48	33	23

Women: Age group
	10–19	20–29	30–39	40–49	50–59	60–69	70+
Sweet	5.55 (5.09, 6.02)	5.66b (5.38, 5.93)	5.22e (4.87, 5.57)	5.09f (4.76, 5.42)	4.55f (4.09, 5.00)	4.40 (3.81, 4.99)	4.57 (3.71, 5.43)
Sour	5.72 (5.24, 6.21)	6.16 (5.93, 6.40)	5.76 (5.45, 6.06)	5.34b (5.03, 5.64)	4.94 (4.48, 5.41)	4.77 (4.21, 5.33)	3.60 (2.70, 4.49)
Bitter	5.30d (4.75, 5.85)	5.42 (5.08, 5.78)	5.26e (4.88, 5.65)	4.81f (4.45, 5.16)	4.55e (4.02, 5.07)	4.00a (3.34, 4.66)	3.81b (2.98, 4.64)
Salty	5.25 (4.73, 5.77)	6.00 (5.76, 6.24)	5.46 (5.14, 5.79)	5.61e (5.32, 5.90)	5.70f (5.33, 6.07)	5.35b (4.87, 5.84)	5.19 (4.47, 5.90)
Umami	2.95 (2.42, 3.48)	3.35 (3.01, 3.69)	2.99 (2.63, 3.35)	2.73 (2.42, 3.03)	2.84 (2.39, 3.30)	3.03 (2.46, 3.60)	2.65 (1.92, 3.37)
Sample size	76	225	186	209	101	65	37

Note: Female test scores that significantly differ from male test scores for a given age group are indicated in bold. See text for details.

p values for statistically significant higher female than male means: ^a≤ 0.05; ^b≤ 0.025; ^c≤0.010; ^d≤0.0025; ^e≤0.005; ^f≤0.001.

3.2. Normative Data Tables for Men and Women

The age‐ and sex‐related normative data useful for determining a subject's general taste function are presented in Table 3. These tables are color coded to provide an indication of function/dysfunction based on age‐ and sex‐corrected norms, making it possible to rapidly determine if a subject's WETT score is normal for his or her age. Scores falling clearly within the normal range are colored in green, while scores falling in the approximate lower third of the distribution, that is, between the 30th and 5th percentiles, are colored in yellow. Scores falling below the fifth percentile, which is a typical percentile for defining abnormality from a normal distribution, are colored in red.

TABLE 3.

Normative data for interpreting male and female WETT scores for the total 53‐item test corrected for subject age.

Note: Scores in green are normal. Those in yellow are ones falling in roughly in the bottom third of the normal distribution, that is, from the 30th percentile to the 5th percentile. Red indicates scores at or below the 5th percentile. See text for details.

It should be emphasized that while the test scores reflect performance relative to age, the absolute performance decreases as age increases and older persons who are normal for their age may still have a significant deficit relative to younger ones. For example, a score of 21 is abnormal for a 10‐ to 19‐year‐old male (3th percentile) but is normal for a man 70 years of age or older (21st percentile). Such percentiles aid in helping patients place into perspective their loss relative to sex‐ and age‐related peers.

Table 4 provides the age‐ and sex‐related percentiles for each of the five basic taste qualities of the WETT subtests. Given the relatively constrained range of test scores, test categories are not color coded in the same manner as those in Table 3.

TABLE 4.

Normative data for interpreting female WETT subscale scores for each of the representative five basic taste qualities. See text for details.

4. Discussion

This paper provides, for the first time, normative data for a unique clinically useful taste test that is self‐administered and requires no liquid stimuli. The tastants are embedded in pads of monomer cellulose located on disposable plastic strips that are self‐applied to the tongue. In addition to detecting differences in age and sex, this test is highly reliable (test‐retest r = 0.92) and is as sensitive, in some cases more sensitive, than classic liquid‐based tests to age, sex, etiology (e.g., head trauma, upper respiratory infections), and the bitter tasting chemical phenylthiocarbamide (PTC) [29]. Self‐administration is efficient and cost effective, eliminates the need for direct physician or technician involvement, and precludes potential examiner influences on test results.

There is a large literature on age‐related taste changes into which our findings fall (for reviews, see [38, 39]). By examining taste test scores for subjects spanning the entire age range, the present study was able to document that a near‐linear age‐related decline occurs in taste function that begins in the 30s and occurs for all basic taste qualities, including umami (Figure 2 and Table 2). Studies that simply compare the test scores of young (e.g., 20–25 years) and old (e.g., >59 years) subjects [38, 40–44] cannot determine when the age‐related decline begins. Our finding that the decline begins relatively early in life contrasts with the belief that taste deficits parallel other age‐related sensory deficits seen after the age of 60 years [38, 45]. This finding begs the question as to what age‐related factors are responsible for the initiation of the decline in taste loss so early in life. Although vallate papillae are more numerous and contain more taste buds in younger than older persons [46], taste bud density of foliate papillae appears to be rather constant throughout life [47]. On the anterior tongue, taste sensitivity is nonetheless directly related to fungiform papilla density [48, 49]. It is important to emphasize, however, that marked individual differences exist in the number of taste buds [50], likely reflecting much of the variability observed in the literature.

The present study clearly demonstrates that women outperform men on a basic taste test. Although some early threshold studies were negative or equivocal on this point (e.g., in a 1959 study of 80 men and 20 women, ranging in age from 15 to 89, no influences of sex was observed) [51], the vast majority of threshold studies have reported lower thresholds (i.e., greater sensitivity) in women than in men (for reviews, see [40, 52, 53]). Greater female performance has also been previously documented for a number of suprathreshold taste tests in addition to the WETT [29, 39, 53, 54].

The basis for the sex differences is not clear, although there is evidence that women have more fungiform papillae and taste pores than men on the anterior tongue [40]. Other potential contributors to sex differences are salivary constituents that may limit the ability of tastants to maintain their integrity or to reach and bind with taste receptors (e.g., salivary enzymes, ions, flow rate, thickness, and volume) [55, 56], hormonal influences on pre‐ and peri‐natal (organizational) brain development [57], and concurrent (activational) differences in circulating levels of gonadal steroids [58]. However, as occurs in olfaction [59], concurrent levels of gonadal hormones do not appear to be correlated in a straight‐forward manner with better female taste performance. For example, inferences of hormonal effects from fluctuations in taste sensitivity across phases of the menstrual cycle vary with the tastant and are not entirely congruent with levels of estrogens. Thus, threshold sensitivity to NaCl is higher in the cycle's preovulatory phase (when estrogens are most elevated) [60], whereas sensitivity to bitter tasting stimuli (quinine, propylthiouracil) is higher during menses when estrogens are lowest [61]. Exogenous estrogens increase, not decrease, the sucrose taste threshold of rats [62], reduce their chorda tympani nerve responsiveness to NaCl [63], and facilitate the depressive effects of glucagon‐like peptide‐1 agonists on the reward value of sucrose in an operant conditioning paradigm [64]. The influences of exogeneous testosterone on taste function are unknown.

Our finding that WETT scores are lower for caffeine and NaCl in smokers than in non‐smokers is in accord with some, but not all, studies on this topic. Thus, no effects of smoking on taste have been reported by a number of investigators [65, 66], whereas others found smoking‐related effects limited to bitter [40, 67], salty [68], or, as in the present study, both bitter and salty tasting substances [69]. These differences likely reflect methodological issues, including sample sizes, psychophysical paradigms, and stimulus concentrations.

The present total WETT norms are likely universally applicable, although subtle cultural differences may exist in responses to its subcomponents. In one study, for example, the overall WETT test scores of 214 American subjects tested in the United States did not differ significantly from those of 113 Chinese participants tested in China (respective means [SEMs] = 35.01 [0.72] and 33.05 [0.98]; p = 0.129) [70]. Although no differences were apparent for sucrose (p = 0.833) or NaCl (p = 0.368), the Chinese scored higher for umami (MSG) (p = 0.024) and the Americans for caffeine (p = 0.029) and citric acid (p = 0.001). The reasons for these differences are not clear, although cultural factors are likely involved. For example, MSG is much more common food additive in China than in the United States, and lower thresholds for MSG have been reported for Chinese subjects [71]. Unlike the United States, coffee, a major source of caffeine taste, is a rare commodity in China [70].

It should be noted that the manner in which the taste strips are sampled can significantly influence the test results. In the current study and other studies from our group, the test strips are repeatedly moved around the edges of the tongue and over its tip—regions where most taste buds are located. When WETT strips are placed only in the center of the tongue (which contains comparatively few taste buds), much lower levels of detection have been reported [72]. For example, in one study of 40 men ranging in age from 20 to 39 years in which this was done, the total WETT mean score of was 28.7. In our study, the mean for the 654 men of this age group was 34.59. Such observations indicate that the WETT is sensitive to how the strips are sampled and that regional tongue differences in sensitivity can be determined employing this test—differences that have been documented for both chemical and electrical taste testing [73, 74].

Our study found no evidence that a prior history of COVID‐19 significantly impacts WETT scores. This lack of an association was observed previously by our group [32] and begs the question as to whether or to what extent taste, per se, is actually impacted by the SARS‐CoV‐2 virus early in the COVID‐19 disease process. We and others employing objective test measures have reported the lack of taste dysfunction during the initial SARS‐CoV‐2 outbreak [30, 75], although consensus on this point is lacking in the literature [76]. Olfaction, which is often confused with taste, exhibits marked COVID‐19‐related dysfunction early in the disease process [77]—dysfunction that continues to some degree in a significant number of afflicted individuals more than a year after the initial infection [32].

The present research has both strengths and weaknesses. Among its strengths are (a) the establishment of normative data of clinical value for a well‐validated self‐administered taste test that employs no liquids and multiple concentrations of representatives of all five basic taste qualities, (b) sampling of healthy subjects proportional in numbers to population frequencies of different regions of the United States, (c) evaluation of key subject variables that could impact taste function, namely, age, gender, COVID‐19 history, and smoking behavior, and (d) provision of evidence that taste function declines in a near‐linear fashion from the third decade of life to old age. Among its limitations are (a) the lack of physical examinations of the participants and the reliance on the participants’ self‐declaration of being healthy, (b) fewer numbers of subjects in the oldest age categories, (c) the absence of taste‐related genetic information on the subjects, and (d) testing limited to the whole mouth rather than to separate lingual regions innervated by CN VII and CN IX. In regard to the latter, it should be noted that regional deficits are extremely rare and whole‐mouth testing is not only practical in terms of patient and examiner time in the clinical setting but is the best reflection of a patient's overall experience of taste function. That being said, WETT taste strips can be employed in cases where damage to specific nerves is of concern.

5. Conclusion

Quantitative taste testing is rarely performed in medical clinics, in part because of the lack of practical means for assessing taste function and the absence of normative data. This study provides clinically useful normative data for interpreting results from the 53‐item WETT, a practical self‐administered taste test that requires no liquid stimuli or rinses. Age‐ and sex‐adjusted percentiles are provided from 1392 healthy subjects sampled from all US states except Hawaii. An age‐related decline in taste function was demonstrated that began in the 30s, as was a sex difference favoring women. Cigarette smokers exhibited lower salty (NaCl) and bitter (caffeine) WETT scores than nonsmokers, although such effects were not apparent for the other taste qualities. This research now makes it practical for clinicians to accurately and quantitatively evaluate the taste function of their patients.

Conflicts of Interest

Richard L. Doty received consulting fees from Johnson & Johnson and Merck, and royalty payments from Cambridge University Press, Elsevier, Johns Hopkins University Press, McGraw‐Hill, and John Wiley & Sons outside the submitted work. He is president and major shareholder of Sensonics International, the manufacturer of the taste test used in this study. Ryan Sharetts and Shima T. Moein are employees of Sensonics International. Participant payments came from Sensonics International's research fund.

Doty R. L., Sharetts R., and Moein S. T., “Self‐Administered Taste Testing Without Water: Normative Data for the 53‐Item Waterless Empirical Taste Test (WETT).” International Forum of Allergy & Rhinology 15, no. 9 (2025): 15, 934–943. 10.1002/alr.23587