When I use a word . . . The languages of medicines—International Nonproprietary Names (INNs)

Alphabets

For the most part, languages have symbols that allow them to be written down. The symbols may be pictographs, single symbols that represent individual syllables, words, or phrases, based on visual representations of entities that the word implies. Japanese, for example, can be written in kana, two forms of syllabic writing, hiragana and katakana, which consist of symbols, adapted from Chinese characters, that represent the syllables of the word. Most languages, however, have words that are built, not from pictures, but from strings of individual symbols that we call letters.

The alphabet in which English is written contains 26 letters, five vowels and 21 consonants. However it contains 44 phonemes in all, i.e. letters or combinations of letters that convey different sounds. So in addition to the basic 26 we have, for example, the phonemes /ch/, /ph/, /sh/, /th/ and /zh/ and the diphthongs /ay-i/, as in “page,” /ee-ə/, as in “here,” and /oh-oo/, as in “slow” (say the words slowly and you’ll hear the diphthongs).

But even when different languages use the same letter symbols they may use different subsets of the total available. Classical Latin, for example, was written using 23 of the letters in the alphabet that we use, omitting the letters /j/, /v/, and /w/. The letter /u/ was used instead of /v/. For example the greeting “farewell,” which we would usually write as “vale” was written “uale” and pronounced as it looks. In addition, the letters /x/ and /z/ were not canonical members of the alphabet; /x/ was used to represent /cs/ and /z/ was imported to represent words of Greek origin that contained the letter zeta.1

In other languages, alphabets include letters based on some of the letters that we use, but extended by diacritics, marks typically placed above letters, inducing a change of pronunciation. In French, for example, the letter /e/ can have different pronunciations depending on the acute and grave diacritics: /é/ and /è/. The letters /e/, /i/, and /o/ can also take a circumflex accent, /ê/, /î/, and /ô/, indicating the loss of the letter /s/ from the Latin word from which the French word is derived. For example, in Latin the word for a window is fenestra, and in French this becomes fenêtre; in Latin island is insula, but in French it is île; and in Latin the word hospes, a guest or host, via the adjectival form, hospitalis, gives us the word hospital, but in French it is hôpital. In each case, the circumflex accent shows the absence of the letter /s/ in the original, which has persisted in English. French also modifies the letter /c/ in some words by the use of a cedilla, written underneath the word, not over it, thus /ç/. This softens the sound in words in which it might be expected to be pronounced hard, changing it from the sound of a kay to that of an ess, as in française.

The umlaut in German serves a similar purpose, modifying the vowels /a/, /o/, and /u/, giving /ä/, /ö/, and /��/. “Umlaut” literally means “around sound.” The change is sometimes marked in English transliterations as /ae/, /oe/, and /ue/; hence, Henoch-Schönlein purpura can be anglicised as Henoch-Schoenlein, as is sometimes done.2 German also features an extra letter, the Eszett, which looks like a long /s/ joined to a /z/, thus /β/, and is pronounced, and often spelt, /ss/. One would be forgiven for confusing it with a Greek beta, /β/.

Spanish and Portuguese have another kind of diacritic, the tilde, a wavy line written above a letter. In Spanish it is used to change the /n/ sound to /ny/, as in cañon, a canyon, and mañana, tomorrow. In Portuguese it indicates nazalisation, as in órgão, organ, and mão, hand.

In contrast to all of these natural languages, and many others, the system of artificial International Nonproprietary Names that are given to medicines, forming a restricted vocabulary with specific rules, uses an alphabet that contains fewer phonemes than the English alphabet allows.

International Nonproprietary Names (INNs)

The system of names known as INNs was introduced by the World Health Organization (WHO) in the 1950s. According to its founding documents, adopted by the International Health Conference, which was held in New York in 1946, ratified by 61 member states, and enforced in 1948, the WHO has a constitutional mandate requiring it to “develop, establish and promote international standards with respect to biological, pharmaceutical and similar products.”3

The first World Health Assembly, in July 1948, called for unification of pharmacopoeias and the preparation of an International Pharmacopoeia, and in 1949 an Expert Committee on Unification of Pharmacopoeias proposed a plan for doing so, including general rules for drug nomenclature; the plan was adopted in 1950, and a further resolution in 1953 established the programme on INNs.4 An expert subcommittee was established in 1950, and the documents “Procedure for the Selection of Recommended International Nonproprietary Names for Pharmaceutical Substances” and the “General Principles for Guidance in Devising International Nonproprietary Names for Pharmaceutical Substances,” were adopted in 1955. Since then the principles have evolved as the types of drug to be named have also evolved.

Assigning an INN

INNs come in two varieties, proposed and recommended.

The process begins when a manufacturer applies for an INN for a new compound, offering several possibilities to choose from. The INN committee reviews the application and offers a name, preferring whenever possible to choose one that the manufacturer has proposed, provided there are no conflicts with other existing names, including brand names, and provided that the choice conforms to a set of stated principles. For example, the name should not imply an unsuitable connotation in any of the major world languages spoken.

A manufacturer may disagree with the committee’s suggested choice, in which case further discussion takes place. When there is final agreement, the agreed name becomes a proposed INN (pINN) and is published as such. After a time, and provided there are no objections, or after any objections have been resolved, the name becomes a recommended INN (rINN) and is published as such, replacing the pINN. If, however, there is an objection that is not resolved, the name remains on the list of pINNs.

Two main principles guide the formulation of INNs:

1. INNs should be distinctive in sound and spelling. They should not be inconveniently long and should not be liable to confusion with names in common use.

2. The INN for a substance belonging to a group of pharmacologically related substances should, where appropriate, show this relationship. Names that are likely to convey to a patient an anatomical, physiological, pathological or therapeutic suggestion should be avoided.

These two principles continue to inform the ways in which INNs are devised. Over the years INNs have tended to become longer and longer, as the number of INNs has grown and it has become increasingly difficult to coin new names that do not conflict with existing names and will therefore not be likely to be confused with them. Efforts are constantly being made to keep new INNs as short as possible and to be easily pronounced by speakers of all the major world languages.

To take an example: the suffix –stat, which is sometimes also used as an infix, -stat-, indicates that the primary pharmacological action of a drug is enzyme inhibition. As the number of such drugs has increased, with the discovery of new enzymes and new drugs to inhibit them, so the number of drugs ending in -stat has increased markedly. This has led to the addition of infixes in some cases, specifying the actual enzyme involved; for example, -castat indicates a dopamine-hydroxylase inhibitor, -elestat an elastase inhibitor, -inostat a histone deacetylase inhibitor, -listat a gastrointestinal lipase inhibitor, -mastat a matrix metalloproteinase inhibitor, -mostat a proteolytic enzyme inhibitor, and -restat an aldose reductase inhibitor; and -vastatin indicates an HMG CoA reductase inhibitor.

The INN alphabet

Proposed and recommended INNs are published in a cumulative list, updated periodically. The names are presented in alphabetical order according to the Latin forms of the names, followed by their names in English, French, Spanish, Arabic, Chinese, and Russian. Because the names are international, the alphabet in which they can be written has to be restricted, and the rules are given in one of the seven secondary principles that follow the two primary principles given above:

7. To facilitate the translation and pronunciation of INNs, “f” should be used instead of “ph,” “t” instead of “th,” “e” instead of “ae” or “'oe,”' and “i”' instead of “y”; the use of the letters “h” and 'k” should be avoided.”

This reduces the number of phonemes available from 44 to 37, which gives an added restriction on choices.

On the whole, these rules are followed, but there are some anomalies. For example, the rINN flupentixol was published in recommended list number 6 in 1966, having first been proposed as such (pINN) in 1964 (proposed list number 14). Although the British Approved Name (BAN) until 1999 was flupenthixol, it was changed to flupentixol when almost all BANs were converted to INNs in 1999, following a 1992 European Community directive, decreeing that in member countries the recommended INN should be used exclusively.5

Nevertheless, zuclopenthixol was listed as the pINN in proposed list number 50 in 1983 and as the rINN in recommended list number 24 in 1984. If the rules had been followed, it should have been zuclopentixol.

There are many other exceptions to the rule that “t” should be used rather than “th,” although most of them are old names that were coined before the rule was introduced, and some of them begin with “th,” such as theophylline, thiopental, and thiotepa. On the other hand, thioguanine was renamed tioguanine.

In an analysis of the formal and semantic properties of 7987 INNs in relation to the WHO's nomenclatural principles naming guidelines of the WHO, I and my colleagues found a range of inconsistencies, such as these.6

A final thought

INNs have gradually been adopted in many countries. For example, they were adopted in Australia in 20167 and in Latvia in April 2020.8 A before-and-after study in Latvia showed that after the adoption of INNs, their use in prescriptions for bisoprolol and/or perindopril increased from 2.1% to 92.3% and the rate of medication errors fell from 0.81% to 0.39%.

No nomenclatural system is perfect, and the INN system is no different from any other in that respect. However, despite inconsistencies, the system has proved useful in bringing a degree of uniformity to the way in which drugs are named internationally.