The bitterness of nine commercial antibiotics (clarithromycin, erythromycin, cefdinil, doxycycline, vancomycin, tetracycline, minocycline, oxytetracycline and bacampicillin) was evaluated in human gustatory sensation tests with nine volunteers. The bitterness of 0.1–0.3 mM solutions (or suspensions in the case of clarithromycin) of the antibiotics was then measured using an artificial multichannel taste sensor. In the sensor measurements, three variables were used to predict estimated bitterness in single and multiple regression analysis and principal component analysis: sensor output as relative value (R), the change of membrane potential caused by adsorption (C) and C/R. Particularly good correlation was obtained between obtained bitterness scores and predicted scores using C from channel 2 of the sensor (r2 = 0.870, P < 0.005) and C/R values for channels 2 and 3 (r2 = 0.947, P < 0.005). The taste sensor was also successful in assessing the bitterness intensity of clarithromycin powder suspensions of various concentrations. Clarithromycin has a low aqueous solubility but is the most bitter of the nine antibiotics. Sensory data from channel 3 of the sensor predicted the bitterness of clarithromycin powder suspensions and their filtered solutions well. Finally, the bitterness intensity of a commercial clarithromycin dry syrup product (Clarith dry syrup, Taisho Pharmaceutical Co. Ltd, Tokyo, Japan) was evaluated in gustatory sensation tests and using the taste sensor. In Clarith dry syrup the drug is coated with aminoalkyl methacrylate polymer using a spray congealing method. The taste sensor results confirmed that the polymer was successful in almost completely masking the bitter taste of the dry syrup product.

We describe a new method for the evaluation of the bitterness of medicines by semi-continuous measurement of adsorption using a multichannel taste sensor or 'electric tongue'. The bitterness of 10 basic medicines was evaluated by both the taste sensor and in human gustatory sensation tests with 11 volunteers. The sensor part of the taste sensor consists of eight electrodes made of lipid/polymer membranes. Three variables were obtained from the taste sensor data: sensor output (S), the change of membrane potential caused by adsorption, corresponding to aftertaste (C), and the ratio C/S. These variables were used to predict an estimated bitterness score in multiple regression analysis. Semi-continuous measurement of C (every 30 s up to 150 s) was adopted as an additional explanatory variable, and the attenuation rate of C was defined as C'. These data were also subjected to multiple regression analysis. The correlation coefficient (r) estimated for the bitterness score predicted by the taste sensor, using C' for channel 2 and C/S for channel 4, and the score obtained by human gustatory sensation, was 0.824. This value was greater than that obtained using C/S for both channels 2 and 4 (0.734). The method described in the present study seems to offer good predictability for the evaluation of bitterness.

The purpose of this study was to evaluate the ability of a quantitative prediction method using a taste sensor to determine the bitterness of clarithromycin powder suspensions of various concentrations and of a commercial clarithromycin dry syrup product (Clarith® dry syrup, Taisho Pharmaceutical Co., Ltd., Tokyo) containing aminoalkyl methacrylate polymer as a taste-masker. The bitterness of the clarithromycin dry syrup product dissolved in various beverages was also evaluated in gustatory sensation tests and using the taste sensor. In the sensor measurements, three variables were used to predict bitterness in single and multiple regression analysis: relative sensor output (R), the change of membrane potential caused by adsorption (CPA), and CPA/R ratio. The CPA values for channel 3 of the sensor predicted well the bitterness of clarithromycin powder suspensions and their filtered solutions. For Clarith® dry syrup, the sensor output was small, suggesting that aminoalkyl methacrylate polymer was successful in almost complete masking of the bitter taste of the dry syrup product. When the bitterness intensities of mixtures of 1 g of Clarith® dry syrup with 25 ml of water, coffee, tea, green tea, cocoa, milk, and a sports drink were examined, a good correlation was obtained between the results from human taste tests and the predicted values calculated on the basis of multiple regression analysis using CPA data from channel 4, and the CPA/R ratio from channel 3 of the taste sensor (r2=0.963, p<0.005). Co-administration of 1 g of Clarith® dry syrup with an acidic sports drink was found to be the most bitter using either method.

The bitterness of 11 commercial medicines was evaluated both by a multichannel taste sensor and in human gustatory sensation tests with 15 volunteers. For basic drugs with amino groups in the molecule, such as quinine, there was a comparatively strong relative response electric potential (mV) of channels 1 or 2, those containing negatively charged membranes and the bitterness determined by human gustatory sensation tests. The suppression of the bitterness of quinine by sucrose and aspartame could be quantified using the artificial taste sensor and the results concurred with those from gustatory sensation tests. The usefulness of the sensor was thus confirmed for this type of compound.Anionic drugs, such as diclofenac sodium or salicylic acid gave rise in a negative response electric potential in channels 5 or 6, those containing positively charged membrane, seemed to be useful information even though their tastes are being sour rather than bitter.For drugs with both an amino (cationic) group and carboxylic acid (anionic) group in the moelcule, such as theophylline, caffeine, and metronidazole, the relative response electric potential (mV) of channels containing negatively charged membranes was not increased, even though bitterness was observed in human gustatory sensation tets. Therefore, a different design of membrane component is required for more general evaluation of the bitterness of various medicines.