Normative data on a reliable and sensitive test of gustatory function

Our sense of taste guides us in our everyday lives. It triggers warning signs, preventing us from ingesting spoiled or toxic substances. It provides us great enjoyment and pleasure, serving as a pivotal element in the multisensory collaboration of flavour within the fundamental act of eating. [1, 2].

As the coronavirus pandemic swept across the world, flavour disturbance followed as the most common symptom of COVID-19 [3]. Medical clinics were suddenly inundated with patients complaining of severe taste disturbances. While this surge brought newfound public attention to taste disturbances, the study of taste function has been a subject of discussion and investigation since at least the time of Aristotle [4]. This interest is rooted in the importance of the sense of taste in our daily lives, although it is often taken for granted. The sweet taste signals energy and induces pleasure-seeking behaviour from birth across species [5]. A bitter taste indicates potential danger but is an acquired taste and has recently been linked with a range of health benefits, from improved digestion and nutrient absorption to stimulating immune function [6].

Thus, taste testing is relevant to researchers and clinicians alike, emphasising the need for reliable measuring methods.

Various gustatory testing methods are currently used, including taste tablets [7], liquid solutions [8] and filter paper [9]. Some testing methods, such as filter paper or taste strips, can yield varied results, possibly because they require a consistent amount of saliva [10]. Furthermore, many existing tests have uneven dilution steps [8], complicating the interpretation of results. This may interfere with proper patient guidance and decision-making concerning optimal diagnosis and treatment. Therefore, it is essential to have a reliable, sensitive and validated method that intuitively links test scores with patient outcomes. In measurements of chemosensory threshold values, the use of turning points is commonplace as it may increase the accuracy of the final threshold score [10]. This is particularly relevant in research, where identifying small changes is essential. A previous study described an easy-to-interpret, high-reliability taste test coined the Taste-Drop-Test (TDT) [10]. However, the TDT contained no normative data on test scores and cut-off values for patients with hypogeusia.

This study aimed to provide normative data and cut-off values for the TDT based on a large sample of patients and normative controls.

Methods

The TDT was applied to 424 healthy participants at Aarhus University and 134 patients from the Taste and Smell Disorder Outpatient Clinic at Regional Hospital Goedstrup [11]. All healthy participants reported no gustatory or olfactory disturbances, while the 134 patients complained of subjective gustatory dysfunction.

Materials

The test consisted of liquid concentrations of four basic tastes: sour, sweet, salty and bitter. The fifth basic tastant, umami, was excluded from the test battery due to the low recognition rate in the general population [12]. Each tastant was stepwise diluted by a factor of two into ten different solutions at the following concentrations:

1. Sour (citric acid) 0.0500 g/ml, 0.0250 g/ml, 0.0125 g/ml, 0.0063 g/ml, 0.0031 g/ml, 0.0016 g/ml, 0.0008 g/ml, 0.0004 g/ml, 0.0002 g/ml, 0.0001 g/ml

2. Sweet (sucrose) 0.5000 g/ml, 0.2500 g/ml, 0.1250 g/ml, 0.0625 g/ml, 0.0313 g/ml, 0.0156 g/ml, 0.0078 g/ml, 0.0039 g/ml, 0.0020 g/ml, 0.0010 g/ml

3. Salt (NaCl) 0.2500 g/ml, 0.1250 g/ml, 0.0625 g/ml, 0.0313 g/ml, 0.0156 g/ml, 0.0078 g/ml 0.0039 g/ml, 0.0020 g/ml, 0.0010 g/ml, 0.0005 g/ml

4. Bitter (quinine hydrochloride) 0.00500 g/ml, 0.00250 g/ml, 0.00125 g/ml, 0.00063 g/ml, 0.00031 g/ml, 0.00016 g/ml, 0.00008 g/ml, 0.00004 g/ml, 0.00002 g/ml, 0.00001 g/ml.

Application

The solutions were first applied at the lowest concentration in a pseudorandomised order between the different basic tastes and water. The solutions were applied with a disposable plastic pipette, and a drop was placed bilaterally approximately three centimetres from the tip of the tongue. The subject was then allowed to close the mouth and attempt to guess the taste in a forced choice paradigm. After each application of the taste solution, the participant rinsed their mouth with water. After each correct identification of a basic taste, the subsequent application of this specific basic taste would be one step lower in concentration (more diluted). After each incorrect guess of the basic taste, the following application would be one step higher in concentration (turning-point application). When a specific basic taste was correctly identified two times for a given concentration, the corresponding concentration step would be noted as the basic taste value ranging from one to ten. The mean of the four basic tastes score was calculated as the TDT score to indicate overall gustatory function (Figure 1). The time for complete application of the test was 12-20 minutes.

Olfactory function

In addition, the 402 healthy participants and the 113 patients with complaints of gustatory dysfunction were tested for olfactory function. The Burghart Messtechnik threshold, discrimination and identification Sniffin’ Sticks (TDI) test were applied. This involves using felt-tipped pens with odorants. In the threshold and discrimination tests, there are sixteen triplets of pens; in the identification test, there are 16 pens. In the threshold test, stepwise diluted odours are presented, and the subject is asked to identify the single pen in each triplet containing an odorant. After two correct guesses, the concentration is lowered; and after one incorrect guess, the concentration is raised. In the discrimination test, the subject is asked to identify the one pen in each triplet with a different odour. In the identification test, each felt tip pen is presented with four answer options, and the subject is asked to choose the correct answer option [13]. The TDI test has been validated in Danish [14].

Statistics

T-tests were performed on comparisons between variables with regression analyses to account for possible confounders. The Wilcoxon rank-sum and Spearman's correlation with Bonferroni adjustment were performed on non-parametric variables. The P value for statistical significance was set at p < 0.05. All analyses were conducted in STATA (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC.).

Trial registration: The Central Denmark Region Ethics Committee approved the study (Reg. No. 42832), which was conducted following the Declaration of Helsinki Ethical Principles of Medical Research.

Results

Healthy participants

The healthy population had a median TDT score of 7.5. Women had a slightly higher median value than men (7.5 versus 7.25). A Spearman’s correlation with Bonferroni correction found a significant positive correlation between female sex and TDT score (p = 0.0004). Furthermore, a significant negative correlation was recorded between higher age and TDT score for both men and women, with a yearly loss of 0.0273 (SE 0.0026) points in TDT score. A Spearman's correlation with Bonferroni correction was made to assess the relationship between TDI and TDT scores using a sample of 402 healthy participants. The TDI and TDT scores had a significant positive correlation (p = 0.0002).

The tenth percentile cut-off for healthy participants was a TDT score of 6.25, which, for both olfactory and gustatory functions, is the threshold for a distinction between normal and reduced chemosensory function [15].

Patients

The population with subjective gustatory dysfunction had a median (interquartile range (IQR)) TDT score of 5 (3.5-6.5).

Spearman’s correlation with Bonferroni correction revealed no relationship between age and patient TDT score (p = 0.15). However, a significant correlation was recorded between female sex and higher TDT scores (p = 0.023). TDI score and TDT score were positively correlated (p = 0.0001).

Among the patients, 36 (27%) had a TDT score above the tenth percentile cut-off value of 6.25. The patients with a TDT score above the cut-off value had a TDI score similar to that of the whole patient group, see Table 1.

Comparison

The data was found to be non-parametric, particularly among the healthy participants. This was due to a ceiling effect, as the mean score of the controls was closer to the lowest dilution, see Figure 2 and Table 2.

The patients scored significantly lower across all four basic tastes than healthy participants did.

Discussion

Main findings

The main finding of this study was the tenth percentile cut-off value of 6.25 to distinguish between normal gustatory function and hypogeusia. We also found an association between age, sex, TDI and TDT score in the healthy control group. These associations align with other studies of chemosensory function [15, 16]. While the age-related yearly taste function decline of 0.0273 points is of little relevance for evaluating the difference of people close in age, the combined effect over several years should interest clinicians, researchers and patients. Surprisingly, age did not affect the TDT score in the patients. Possibly, the gustatory dysfunction overshadowed the effect of age.

The sense of smell and taste intuitively seem associated, as they are both chemosensory-based and involved in food intake and enjoyment. Therefore, in line with recent studies [16], this study found a statistically significant positive correlation between TDI and TDT scores in patients and healthy participants.

A total of 36 (27%) patients with subjective taste loss had no measurable ageusia or hypogeusia. The subjective gustatory function is famously inaccurate compared to the measured function due to a misinterpretation of the olfactory input [17]. In this study, 50% of the patients with normal gustatory function and an olfactory impairment classified their olfactory impairment as a subjective taste disorder. Similar studies consistently rated patient assessments of subjective gustatory dysfunction as uncertain [18, 19]. The median TDI score in the 36 patients without measurable taste loss was similar to the median TDI score of the whole patient group. However, median TDI scores in both patient groups were below the cut-off value for a normosmic olfactory score [14]. A potential hypothesis is that the 36 patients had diminished abilities in distinguishing between the two chemosensory inputs. However, the ability to distinguish unisensory flavour input falls beyond the scope of this study.

Several studies evaluate different testing methods for gustatory function [7, 8]. To our knowledge, however, no studies have evaluated a taste test by how well it distinguishes self-identified patients with gustatory dysfunction from self-identified healthy participants. The TDT reclassifies 27% of the patients from subjective gustatory dysfunction into measured normogeusia. Although we found a strong association between subjective complaints and measured function, the unreliable subjective assessment of taste function highlights the need for clinical gustatory testing in patients with subjective chemosensory complaints. Any measure of gustatory function should take into account the complexity of flavour perception and the patients’ understanding of taste and smell.

Strengths and limitations

The stepwise dilution of the basic tastes is a major strength of the TDT. This enables straightforward interpretation of the test results because each point on the ten-point scale has the same value. Thus, a patient presenting with a four-point loss in TDT score has a loss that is four times as high as someone presenting with a one-point loss. This makes comparisons between patients from different procedures, aetiologies or age groups very intuitive and understandable for statistical and communicative purposes. Another strength is the high test-retest reliability [10]. While the test is slower than some alternatives [10, 20], the turning-point application and tester-controlled droplet method make for precise measurements. This is because each final value has been identified three times, and each application of the solution is highly controlled compared to the test where the subjects drink themselves. Using droplets instead of dry powder or taste strips enables the identification of gustatory dysfunction in patients with reduced saliva production. Other strengths of this study are the large number of healthy participants and the definition of cut-off values, which distinguishes between normogeusia and hypogeusia.

A potential limitation is the application of whole-mouth testing, which rules outperforming unilateral testing. Another limitation is the relatively low number of patients compared with controls.

Consequently, we cannot comment on the usefulness of TDT in distinguishing between patient subgroups. All three types of subjective dysgeusia (parageusia, hypogeusia and ageusia) had a reduced TDT score but were not significantly different from each other. Furthermore, the data on the healthy participants was non-parametric, as some participants reached a ceiling effect. Future studies are needed to reproduce the findings with additional dilution steps to avoid ceiling effects.

Conclusions

The TDT is a stepwise-diluted, turning-point applicated taste test with a high test-re-test reliability. It can distinguish between normo- and hypogeusia. We suggest a cut-off score of 6.25. In healthy participants, taste function is positively associated with age, sex and olfactory function. The TDT can be used by researchers and clinicians when an accurate measurement of gustatory function is warranted.