(M1330-03-18) Can Individual Taste Perceptions Bias the Results of Taste Panel Studies? Lessons Learned from an In Vivo Investigation with Paediatric ODMTs

Purpose: Although many efforts have been made to enable children access to safe and effective medicines, there is still a lack of age-appropriate paediatric formulations for a number of drugs listed in the WHO Model List of Essential Medicines for Children (EMLC). As an approach to improve this situation, a novel paediatric platform concept based on orodispersible mini-tablets (ODMTs) using a combination of flavours and sweeteners as taste-masking strategy was successfully developed for several WHO EMLC compounds in a previous study [1]. As required for paediatric use, these ODMTs are based on safe excipients and offer high dose flexibility and good swallowability. However, evidence of adequate taste and mouthfeel has been lacking. As in vivo studies with paediatric patients present several challenges, taste assessment studies with adult taste panels can provide first indications for the selection of suitable flavours or flavour mixtures. However, even with these adult taste panels, inter-individual differences in taste perception, such as a dislike of certain flavours or flavour combinations, can significantly influence the results of these taste evaluations. As a result, potentially suitable formulation candidates run the risk of not being pursued further due to inadequate demonstration of taste-masking properties. The aim of the present study was to test the taste-masking strategy of the developed furosemide- (FUR) and hydrochlorothiazide- (HCT) ODMTs in an in vivo study. The inclusion of identically-flavoured placebo formulations as an additional measure was intended to facilitate a proper interpretation of the results, taking into account the individual likes or dislikes of the study subjects.

Methods: Taste-masked and non-taste-masked ODMT formulations containing FUR and HCT as active ingredients and placebo formulations were produced on a rotary type tablet press (Riva Piccola, Argentina) using 5 mm biconvex punches. After manufacture the ODMTs were subjected to analytical testing according to Ph. Eur. requirements to ensure product quality for the subsequent in vivo study. The in vivo taste study was conducted on an adult taste panel comprising 12 healthy volunteers (7 females / 5 males, age: 18-36 years). The study followed a cross-over, single-blinded design and was approved by the local ethics committee. At baseline, all subjects were tested for their ability to perceive a bitter taste by rating a quinine hydrochloride solution according to the bitterness value test of the Ph. Eur. (2.8.15). On each study day, volunteers were instructed to refrein from using oral hygiene products for 2 h prior to the test and to refrain from consuming food and beverages (except water) for 1 h prior to the test. To evaluate the taste of the ODMTs, a single tablet was placed on the centre of the tongue, and the volunteers recorded their taste perceptions, including bitterness, mouthfeel, and overall impression, after 15 s and 45 s. After 90 s, the volunteers rinsed their mouths with water, and after a total duration of 3 min, they provided a final evaluation of the aftertaste (Figure 1). At each specific time point, the volunteers' taste perception was recorded using visual analogue scale (VAS) questionnaires. At the end of the 3-min test period, any ODMT residues were collected and the oral cavity was rinsed with water to prevent systemic drug absorption. The taste-masking properties of the formulations were assessed by comparing the taste impressions of the taste-masked and non-taste-masked ODMTs using a Wilcoxon signed-rank test.

Results: The taste-masked and non-taste-masked verum- and placebo ODMT formulations were successfully manufactured and met pharmacopoeial requirements. All volunteers enrolled in the study were able to perceive the bitter taste according to the Ph.Eur. bitterness value test and completed the in vivo study without any adverse effects. The results of the study indicate that the chosen taste-masking strategy is suitable for masking the poor taste of both APIs. However, when comparing the success of masking a poor taste, particularly in terms of bitterness perception, only the taste-masked HCT-ODMTs were rated significantly better than their non-taste-masked counterparts at all taste assessment time points (Figure 2, grey boxes and whiskers). This was not the case for the FUR-ODMT formulations, which only showed a clear trend at the first taste rating time point (15 s). However, although this difference was statistically significant, the taste perceptions of the non-taste-masked FUR-ODMTs were characterised by a high inter-subject variability. When considering individual taste perceptions of placebo formulations, FUR-ODMT results can be interpreted differently. Examples of individual taste perceptions considering the placebo-, FUR- and HCT-ODMTs are shown in Figure 2 (open and filled symbols).

Conclusion: The results of the present study highlight the challenges posed by inter-individual taste preferences in in vivo taste assessment studies. Furthermore, the results emphasise that the inclusion of placebo formulations can serve as a valuable control to confirm that the observed taste perceptions are due to the poor tasting drug and not to individual dislike of the chosen flavour (mixture).

References: 1. 1. Freerks L, Sommerfeldt J, Löper PC, Klein S. Safe, swallowable and palatable paediatric mini-tablet formulations for a WHO model list of essential medicines for children compound - A promising starting point for future PUMA applications. Eur J Pharm Biopharm. 156 11-19 (2020)


Figure 1: Sequence of an in vivo taste evaluation of the orodispersible mini-tablets (ODMTs).


Figure 2: Visual analogue scale (VAS) scores for bitterness perception of three exemplary individual volunteers after taste evaluation of the non-taste-masked (open symbols) and the taste-masked (filled symbols) ODMT formulations; box-whisker plots in light grey represent the range, 25/75% quartiles and median values of all volunteers, VAS score from 0 = very poor impression to 100 = very good impression, * - significant difference (p < 0.025), n = 12.