日本語版

Topic hub — 2026-09-19

Handling the bitterness of medicines

Bitterness in food has reasons to be removed and reasons to be kept. In medicines there is almost no reason to keep it. So this field organised its ways of measuring and masking bitterness earlier than food did: applying compounds directly to the tongue for detection thresholds, putting numbers on them with taste sensors, coating them so they never reach the mouth at all.

Even so the story does not stay simple. The act of masking can produce a different unpleasantness, and the premise that bitter medicines go untaken has, within what was read here, no established quantitative backing. What follows goes in order: measuring, masking, and what masking costs.

What this page establishes

• When 9 cardiovascular medications were applied directly to the human tongue, all were predominantly bitter, with detection thresholds from 0.048 mM to 0.438 mM.
• Masking divides broadly in two: interfering with taste transduction, and preventing release into the mouth at all. Cyclodextrin inclusion belongs to the second.
• How far it falls varies widely by method. Clarithromycin dry syrup is estimated at about 1% of the bitterness intensity, while a 1:2 propranolol-cyclodextrin system reaches only about a 30% reduction on a taste sensor.
• Over-masking works against itself. In orally disintegrating tablets with 5% and 10% citric acid, overall palatability scores fell to 26 and 20, below the 38–46 of the unmodified equivalent.
• WHO and EMA both state that palatability can affect acceptability and adherence, and neither guidance reports an effect size. The quantitative link is not established by measurement in the literature read here.

How the bitterness of medicines is measured

The most direct method is to put it on the tongue. When 9 cardiovascular medications including propranolol hydrochloride were applied to the tongues of human volunteers, all were predominantly bitter, and detection thresholds across the 9 ranged from 0.048 mM (propafenone) to 0.438 mM (procainamide).

If the tongue is the instrument, vocabulary comes first. The most common confusion of terms in the sensory evaluation of medicinal products is between actual bitterness and similar sensations such as sourness and astringency. Choosing a vocabulary the panellists understand is stated as a precondition of the measurement.

This is the same problem as on the food side, and the reason this site keepsastringent and bitter andsour and bitter apart as crossings.

Masking divides into two kinds

A systematic review of cyclodextrins divides taste-masking strategies into two groups. The first interferes with taste transduction, using flavour chemicals and bitterness inhibitors that compete with the drug in exciting taste receptor cells. The second prevents release of the unpleasant-tasting drug into the oral cavity, using ion-exchange resins, solid dispersions, polymer coatings, prodrugs, microcapsules, liposomes, nanoemulsions and cyclodextrin inclusion complexes.

The division is between suppressing after the tongue registers it and never letting it reach the tongue. Almost all the cases below belong to the second.

How far it falls

Start with the largest reduction. When clarithromycin, known for its severe bitterness, is made into a dry syrup by physical coating, the sensory bitterness intensity is estimated to fall to about 1% of that of an equivalent powder suspension, supported both by human gustatory evaluation and by taste sensor prediction. This comes second-hand through a review, and the panel size and method of the original could not be confirmed.

The same drug has been handled another way. Prepared as ethylcellulose microcapsules by emulsion-solvent diffusion and made into a dry suspension, it sufficiently alleviated the bitterness of the commercial product on a 3-level evaluation by 9 healthy volunteers. No statistical test is reported.

With quinine, the degree of masking is shown on a different yardstick: dissolution. Pellets prepared for paediatric use by extrusion-spheronisation and coated with 10%, 20% and 30% Eudragit EPO released 9.2%, 5.9% and 2.1% of the dose in the first 5 minutes, corresponding to lower bitterness scores on the taste sensor. The 20% coating was judged optimal. How much reaches the tongue is itself the index of bitterness.

Inclusion gives a smaller reduction. A 1:1 propranolol to hydroxypropyl-beta-cyclodextrin system scored 8.5 for total bitterness (a 15% reduction) and a 1:2 system 7.3 (about 30%). This was a sensor-only measurement with no human panel, and the anchors of 0 and 10 were assigned rather than measured against a bitterness standard.

The same idea appears on the herbal side. In a decoction of a paediatric formulation containing Platycodon root, combining 6.5% HP-beta-cyclodextrin with 10% xylitol turned the assessment from "difficult to accept" to "sweet with no lingering bitterness": inclusion and sweetness used together, one hiding and one covering.

What masking costs

Suppressing bitterness does not necessarily improve the overall impression. Varying citric acid in orally disintegrating tablets containing a bitter drug, the overall palatability score during disintegration fell from the 38–46 of the unmodified equivalent to 26 at 5% and 20 at 10%. The bitterness was suppressed while the high concentration of citric acid itself damaged the overall impression.

The same study also contains the opposite result. In the evaluation after expectoration, overall palatability with 0.1% and 2.5% citric acid (54 and 55) was higher than without it (44). There is a window of concentration, and the direction also changes with when the evaluation is made. This is one drug substance in one dosage form.

Why it is treated as a problem

The premise is that a medicine hard to take will not be taken. Both the WHO and EMA guidance state that patient acceptability is likely to affect adherence and through it the safety and efficacy of the medicine, and both list palatability and ease of swallowing first among the product characteristics that determine acceptability.

That card carries measurement alongside the position statements. Of 153 children aged 3 to 10 who responded, 89 (58.2%) reported having refused a medicine, and of the 86 who gave a reason, 84.9% (73/86) gave taste or flavour. Swallowing or choking accounted for 8.1% (7/86).

At the same time, the last piece of evidence on that card states its own limit. How many of the events where medication becomes suboptimal are due to taste alone is unknown, and the authors say prospective studies are needed. Within what was read here, the quantitative link between palatability and adherence is not established by measurement. The position statements and the measurements have to be read apart.

The regulatory requirement itself is concrete. Information on the palatability of the active substance should be obtained early, and unless otherwise justified the palatability of a paediatric formulation should be satisfactory in itself, that is, without mixing it into food or drink. Meanwhile the guideline gives no acceptance criteria, no standard method and no thresholds.

Why children matter here is recorded too. In dosage forms where the active substance can interact with the taste buds, taste is one of the major determinants of product acceptability, and the review states that children are more sensitive than adults to bitterness, which unlike sweetness is inherently aversive. That sentence is itself a statement inside a review, and the papers it cites have not been read.

Questions people ask

Can the bitterness of a medicine be measured by machine?

Taste sensors are in use, but they have not replaced human evaluation. The propranolol case was sensor-only with no human panel, and its anchors of 0 and 10 were assigned rather than measured against a bitterness standard. In the quinine case, the sensor's bitterness scores are presented as corresponding to dissolution in the first 5 minutes. The machine's numbers become readable when they are tied to release or to formulation.

How far can bitterness be removed?

It depends on the method. Clarithromycin dry syrup is estimated to fall to about 1% of the sensory bitterness intensity, while a 1:2 propranolol-cyclodextrin system reaches about a 30% reduction on the sensor. With quinine, a 30% coating cuts release in the first 5 minutes to 2.1%, yet the coating judged optimal was 20%. Removing it entirely is not always the goal.

Should more bitterness-suppressing additive simply be used?

There is a case where more worked against it. With 5% and 10% citric acid in orally disintegrating tablets, overall palatability fell to 26 and 20 against the 38–46 of the unmodified equivalent. At 0.1% and 2.5%, by contrast, the scores after expectoration (54 and 55) were higher than without it (44). There is a window, and the direction shifts with when the evaluation is made.

Do bitter medicines really go untaken?

Of 153 children aged 3 to 10 who responded, 89 (58.2%) reported having refused a medicine, and of the 86 who gave a reason, 84.9% gave taste or flavour. But these are self-reported reasons for past refusals, and adherence itself was not measured. Neither the WHO nor the EMA position statement reports an effect size.

What is not known

• The quantitative link between palatability and adherence is not established by measurement in the literature read here. The guidance stops at a qualitative statement that it is likely to have an effect, without an effect size.
• The estimate of about 1% for clarithromycin comes second-hand through a review; the panel size and method of the original could not be confirmed.
• Taste sensor numbers cannot be compared with each other. The 15 to 30% for propranolol and the about 80% cited from an EMA report differ in formulation, in scoring convention and in reference state.
• The relation between citric acid concentration and overall palatability comes from one drug substance in one dosage form. Whether the window sits in the same place for other drugs is not known.
• The statement that children are more sensitive to bitterness than adults is made inside a review, and the papers it cites have not been read.
• The guideline gives no acceptance criteria, no standard method and no thresholds for palatability. What counts as satisfactory is not settled from the document side.

Read in the encyclopedia

• Pharmaceutical taste masking — the classification of methods and the drug-by-drug cases, gathered on the process entry
• Palatability (medicinal product) — the regulatory definition and position statements, and the confusion of terms
• Clarithromycin — known for severe bitterness; two different masking routes sit side by side
• Propranolol — measured both by direct application to the tongue and by cyclodextrin inclusion
• Quinine — the reference compound for bitterness, appearing here as the thing being hidden
• Platycodon grandiflorus — the herbal side, where inclusion and sweetness were combined
• Cyclodextrins — the compound class that keeps the drug from reaching the tongue

Go deeper in the research notes

• Medicine, bitterness and palatability — the distance between the regulatory documents and the formulation research
• The pharmacopoeia and the monastery — the historical side of getting bitter medicine taken
• What the bitter receptors divide between them — why so many drugs taste bitter, from the receptor side
• Tonic water and quinine — how bitterness is handled once a medicine becomes a drink

Source ledger

The primary sources and limitations for the claims on this page can be checked from the source links inside each card. The full list is on the Japanesesource ledger.