The Brompton Cocktail: 19th century origins to 20th century demise – by David Clark

Published on: Author: David Clark 6 Comments

Modern day pain specialists continue to be fascinated by the actions and interactions of particular drugs in specific combinations.   As the art and science of such work progresses it is worth reminding ourselves of practices and assumptions that prevailed in the not too distant past – and how quickly these could change. In this respect, the story of the rise and demise of the Brompton Cocktail has a lot to tell us. I detail it here as an interesting vignette within the history of hospice and palliative care, paying particular attention to the important work of Dr Robert Twycross (pictured).  In the coming period I am hoping to apply the same approach to other key elements of hospice and palliative care practice.


Cicely Saunders (1958) praised the ‘Brompton mixture’ in her first ever paper, written as a medical student and published in 1958. Using as her source Richardson and Baker’s The Management of Terminal Disease in the Practice of Medicine (which had appeared in 1956) she described the ingredients as:

  • nepenthe or liquor morphini hydrochloride
  • cocaine hydrochloride
  • tincture of cannabis
  • gin
  • syrup
  • chloroform water

Such mixtures had become established in Britain’s hospitals during the inter-war years. They can be traced back to the particular combination of morphia and cocaine in used in the late nineteenth century by Herbert Snow, Surgeon to the Cancer Hospital (later the Royal Marsden), situated in Brompton, London. Among his various writings, two 1890  pamphlets are important: ‘On The Re-appearance (Recurrence) of Cancer After Apparent Extirpation’  and ‘The Palliative Treatment of Incurable Cancer, with an Appendix on the Use of the Opium Pipe’ (Snow 1890a,b). In a paper published in the British Medical Journal in September 1896, he enlarged on this latter topic, asserting that malignant disease had its causation in neurosis and arguing that morphia and cocaine, when conjoined, could arrest the progression of the disease and produce ‘conspicuous improvement, sometimes to a marvellous degree’ (Snow 1896:718). Less than a year later, however in a letter to the BMJ Snow complained that ‘for imperative reasons of hospital finance I have been reluctantly compelled to abandon this costly medicine [cocaine] for the majority of my hospital patients’ (Snow 1897:1019).

It appears that these ideas about combining morphia and cocaine crossed the road to the Brompton Hospital. One writer has suggested that it was surgeons there in the 1920s such as Arthur Tudor-Edwards and JEH Roberts who took the notion and applied it to their post-thoracotomy patients (Kerrane 1975), making the combination of the two drugs palatable in a base of gin and honey. Clifford Hoyle, the chest physician, has also been attributed with its development and for its introduction at King’s College Hospital in the 1930s (Hinton 1996, interview). Cicely Saunders (1993) likewise records that it was in use by 1935 as part of a regime of ‘regular giving’ for pain relief at St Luke’s Hospital, Bayswater.

By the middle of the century it appears that such mixtures had become widely adopted and were made available for the patient to drink on demand or at regular intervals. In 1952 the Brompton Hospital produced its own supplement to the National Formulary and the mixture appeared in print for the first time, under the name Haustus E. (Haustus meaning a draught or potion, and E. perhaps elixir).

  • Morphine hydrochloride ¼ grain
  • Cocaine hydrochloride 1/6 grain
  • Alcohol 90% 30 minims
  • Syrup 60 minims
  • Chloroform water to ½ fl. oz.

This version was then listed in Martindale’s Extra Pharmacopoeia in 1958. In 1976 it had appeared in the British National Formulary and gradually it had come to be known by several different names: Brompton cocktail; Brompton mixture; Mistura euphoriens; Mistura pro moribunda; Mistura pro euthanasia; Saunders’ mixture; Hoyle’s mixture. Indeed when the first modern hospices opened in the United Kingdom, they too adopted their own nomenclatures, as at Sheffield where there was St Luke’s mild; St Luke’s moderate followed by (rather coyly) St Luke’s individual (Noble, personal communication 2001).

As the use of the mixture proliferated, so its contents became more variable. The alcohol could take the form of gin, whisky or brandy and might be included in various quantities according to preference. A phenothiazine began to be added, either prochlorperazine or chlorpromazine, for both anti-emetic and sedative purposes. Further afield, in France, there is evidence that an antihistamine, such as promethazine was also being included, in what were there called ‘lytic cocktails’ (Meunier-Cartal, Souberbielle and Boureau 1995). Above all, some practitioners in the United Kingdom favoured diamorphine over morphine; some even dropped the cocaine and used morphine and diamorphine together in the mixture (Lancet 1979:1220).

Scrutinising the cocktail

Enter Dr Robert Twycross.

He had first met Cicely Saunders in 1963. Indeed the following year, whilst still an undergraduate he had created the Radcliffe Christian Medical Society simply in order to give a pretext to invite her to speak in Oxford. As he once observed in an interview with me ‘That meant I went into her black book as a doctor, or a future doctor, who might well be interested in hospice care’ (Twycross 1996, interview). So it was that in 1968 an invitation duly came to join her team at the newly opened St Christopher’s. Robert declined the offer, however, in favour of completing his MRCP, and it was not until 1971 that he finally went as a Clinical Research Fellow to the hospice. There, continuing the early research initiated by the late Ron Welldon, he opened up the Brompton Cocktail to unparalleled clinical and scientific scrutiny.

Over the next few years his work focussed on a number of areas: standardisation of the mixture; the relationship between the active constituents and the vehicle; the keeping properties of the mixture; the role of cocaine within it; and also the relative efficacy of the morphine and diamorphine. Indeed, between 1972 and 1979, Twycross produced 39 publications on these and related themes.

In a 1973 paper ‘Stumbling blocks in the study of diamorphine’, which appeared in the May issue of the Postgraduate Medical Journal, he reported on the limited shelf life of the drug in solution, its potency ratio vis a vis morphine, the lack of research into its oral administration, the insensitivity of current assays, the determinants of undesirable side effects, and the importance of between sex comparisons of metabolic handling (Twycross 1973). The following year, in a paper written with colleagues from the Epsom Hospital Laboratories, he began to advance the case for the oral administration of strong narcotics, demonstrating from a study of urinary excretion that ‘an orally administered solution of diamorphine hydrochloride is completely absorbed by the gastro-intestinal tract but that a solution of morphine sulphate is only two-thirds absorbed’ (Twycross, Fry and Wills 1974:493). This difference, he suggested, could be allowed for in the dosage.

Also in 1974, Twycross reported on a survey of 90 teaching and district general hospitals in the United Kingdom which showed marked variation in the composition of what he was now calling ‘elixirs’ for the relief of pain and suffering in terminal cancer. He welcomed therefore the introduction of a standard diamorphine and cocaine mixture to the British Pharmaceutical Codex, but raised questions about its keeping properties. Another issue concerned the acceptability to patients of a mixture which might be experienced as either extremely sickly or unacceptably alcoholic. Above all, the paper pointed to the need to ‘evaluate objectively the contribution of the cocaine to the pharmacological effect of the mixture’ (Twycross 1974:159).

Transatlantic perspectives

Across the Atlantic, others were being influenced by the British hospice movement’s adoption of the Brompton mixture as well as the notion of ‘total pain’ which had been developed by Cicely Saunders. In Canada, Balfour Mount, Robert Melzack and colleagues began a series of studies at the newly opened palliative care unit in Montreal. Early descriptions suggested that ‘the Brompton mixture provides convenient and uniform pain control without important adverse effects’ (Mount, Ajemian and Scott 1976:124). It was found that the mixture relieved the pain of 90% of patients in the palliative care unit and 75-80% of those in the general wards and private rooms, an interesting side-light on the added benefits of the palliative care setting. It was concluded that the results were consistent with the gate control theory of pain and that the Brompton mixture ‘does not act on only a single dimension of pain but has a strong effect on the sensory, affective and evaluative dimensions together’ (Melzack, Ofiesh and Mount 1976:128). So far, it seemed, the traditional elixir, albeit with greater specificity as to its makeup, had survived the transition from old-style care of the dying, to the new world of palliative care.

A key finding

Then an important breakthrough came in two papers published by Robert Twycross in the year 1977. The first appeared in the journal Pain (Twycross 1977a). Here a controlled trial of diamorphine and morphine was reported in which the two drugs were administered regularly in a version of the Brompton mixture containing cocaine hydrochloride in a 10mg dose. A total of 699 patients entered the trial and of these 146 crossed over after about two weeks from diamorphine to morphine, or vice-versa. The previously determined potency ratio of 1.5:1 was used. In the female crossover patients, no difference was noted in relation to pain or other symptoms evaluated, but male crossover patients experienced more pain and were more depressed while receiving diamorphine, suggesting that the potency ratio was lower than expected. Robert concluded that if this difference in potency is allowed for, then morphine is a satisfactory substitute for orally administered diamorphine; but that the more soluble diamorphine retained certain advantages when injections are required and doses are high.

In the second trial (Twycross 1977b), which was reported in a letter to the BMJ, the morphine and diamorphine elixirs were compared with and without cocaine. There were 45 satisfactory crossovers and since the trends within the morphine and diamorphine groups were similar, they were combined for purposes of analysis. The study showed that introducing a 10mg dose of cocaine after two weeks resulted in a small, but statistically significant difference in alertness; but stopping cocaine after this period had no detectable effect. Robert adjudged that at this dosage, cocaine is of borderline efficacy and that tolerance to it develops within a few days.

As a result of this work, the routine use of cocaine with patients at St Christopher’s was abandoned and in particular the morphine was prescribed alone in chloroform water, together with an anti-emetic, where indicated. As Cicely Saunders (2000 HHP interview) put it in an interview with me last year: ‘So one day in May 1977 all the patients who were on oral diamorphine went onto oral morphine and the only people who turned a hair were the doctors who had to write out all the drug charts.’

Supporting evidence came two years later from the Canadian researchers, who also reported a double-blind crossover in which a standard Brompton ‘cocktail’ containing morphine, cocaine, ethyl alcohol, syrup and chloroform water was compared with morphine alone in a flavoured water solution. Pain was measured using the then recently developed McGill Pain Questionnaire and ratings of confusion, nausea and drowsiness were obtained from the patients, their nurses and relatives. The study showed no significant difference between the ‘cocktail’ and the oral morphine alone; both relieved pain in about 85% of patients, with no differences in confusion, nausea and drowsiness. The Canadians adopted the name ‘elixir of morphine’ for the morphine solution (Melzack, Mount and Gordon 1979).

Imminent demise

The demise of the Brompton mixture was now drawing near – at least among the experts.

In 1979, as one of three chapters he wrote for the important trilogy Advances in Pain Research and Therapy, edited by John Bonica and Vittorio Ventafridda, Robert drew together his summative statement on the matter (Twycross 1979). There had been, he suggested, a tendency to ‘endow the Brompton Cocktail with almost mystical properties and to regard it as the panacea for terminal cancer pain’ (1979:291-2). Generously, he allowed that if the physician is aware of the potential side effects of the main ingredients, then its use might be maintained. But set against this was the disadvantage to the pharmacist, the potential unpalatability to the patient, the higher financial costs incurred, and the restricted potential for the physician to manipulate the doses given. The Brompton Cocktail, it turned out, was no more that a dressed-up way of administering oral morphine to cancer patients in pain. It was about to depart from the received wisdom of the new palliative care movement.

Despite such statements, the Brompton Cocktail lived on for some while. Often referred to as Brompton’s Mixture in North America, it continued to be endorsed there in the professional literature (Glover et al 1980). Interviews conducted as part of the Hospice History Project at the University of Sheffield suggest it stayed in use in community services and in British hospitals during the 1980s. Work in progress at Sheffield indicates that it is still known to older people living today. It flowered briefly in the early days of the modern hospice movement, which encouraged its rise, but which through the work of Robert Twycross quite quickly also produced its demise.

What are the implications of this interesting interlude in the history of pain control? I’ll close with some more general reflections.


It is not difficult to see the fall of the Brompton Cocktail as part of a wider sea-change in the science and art of terminal care. We might capture this as a shift from a ‘traditional’ mode of thinking and practice to one which is distinctly ‘modern’ in character. Potions, mixtures, and above all elixirs carry ancient associations, reaching back to earlier periods in the history of medical practice. They can be invested with mystical, even alchemical properties. Yet their actions can be as clouded as their appearance. The conjoining of substances and liquids has an intuitive, even hubristic quality – a medicine of faith rather than fact. At the same time, as the varying names of this particular mixture reveal, the purposes of its use were also somewhat ambiguous. Was it intended to induce euphoria? Did it respond to, or propentiate, moribundity? Most intriguing of all – was it intended to bring about an ‘easeful’ death or actually to hasten demise?

There seems to be no doubt that the Brompton Cocktail was spoken about euphemistically by doctors and nurses – and this was picked up by patients and families. Such ways of communicating are hard to abandon – witness the McGill term ‘morphine elixir’ for a mixture that contained only the drug itself and water. But in general this has not continued.

Instead, an allegedly more rational approach has been adopted to the management of pain since the demise of the Brompton Cocktail. Yes, drugs are used in combination (often in palliative care for purposes for which they are not licensed) but titration has become the important watchword – something much harder to do with the varied ingredients of the cocktail.

Encouraged by Cicely Saunders at St Christopher’s Hospice and mentored by Duncan Vere at the London Hospital, Robert Twycross saw the implications of all this. In the current sociological parlance he ‘deconstructed’ the Brompton Cocktail and revealed it as several things: as a set of pharmacological problems; as a practical matter of manufacturing and safe storage; as an issue involving patient preferences; as a symbolic element in the culture of terminal care; and above all as a key determinant of future thinking and practice about pain control. As Robert’s work was disseminated it became clear that simpler, more predictable means of pain control could be adopted, that narcotics could be used safely – in particular, that morphine was just as effective as diamorphine. Combining energies from Britain, North America and Europe a new field of pain medicine and research was opening up – a field with which the particular contribution of Dr Robert Twycross will always be associated.

David Clark


For a full account see: Clark D (2003) The rise and demise of the “Brompton Cocktail” In ML Meldrum (ed) Opioids and Pain Relief: a Historical perspective. Progress in Pain Research and Management, Vol 25. Seattle: IASP Press, 85-98.


Glover DD, Lowry TJ and Jacknowitz AI (1980) Brompton’s mixture in alleviating pain of terminal neoplastic disease. Southern Medical Journal, 73(3): 278-82.

Hinton J (1996) Hospice History Project interview with David Clark.

Kerrane TA (1975) The Brompton Cocktail. Nursing Mirror, May 1st: 59.

The Lancet (1979) Editorial. The Brompton Cocktail. 9 June: 1220-1.

Melzack R, Ofiesh JG and Mount BM (1976) The Brompton mixture: effects on pain in cancer patients. Canadian Medical Association Journal, July 17, Vol 115: 125-9.

Melzack R, Mount BM and Gordon JM (1979) The Brompton mixture versus morphine solution given orally: effects on pain. Canadian Medical Association Journal, 17 February, Vol 20: 435-8.

Meunier-Cartal RN, Souberbielle JC and Boureau F (1995) Morphine and the “Lytic Cocktail” for terminally ill patients in a French general hospital: evidence for an inverse relationship. Journal of Pain and Symptom Management, 10(4): 267-73.

Mount BM, Ajemian I and Scott JF (1976) Use of the Brompton mixture in treating the chronic pain of malignant disease. Canadian Medical Association Journal, July 17, Vol 115: 122-4.

Noble, Bill (2001) Personal communication.

Richardson JS and Baker D (1956) The Management of Terminal Disease. In the Practice of Medicine. London: Churchill.

Saunders C (1958) Dying of cancer. St Thomas’s Hospital Gazette, 56(2): 37-47.

Saunders C (1993) Oxford Textbook of Palliative Medicine. Oxford: Oxford University Press, pp v-viii.

Saunders C (2000) Hospice History Project interview with David Clark.

Snow H (1890a) The Palliative Treatment of Incurable Cancer: With an Appendix on the Use of the Opium Pipe. London: J and A Churchill.

Snow H (1890b) On the Re-appearance (Recurrence) of Cancer After Apparent Extirpation, with Suggestions for its Prevention. London: J and A Churchill.

Snow H (1896) Opium and cocaine in the treatment of cancerous disease. British Medical Journal, September 19: 718-9.

Snow H (1897) The opium-cocaine treatment of malignant disease. Letter. British Medical Journal, April 17: 1019.

Twycross RG (1973) Stumbling blocks in the study of morphine. Postgraduate Medical Journal, May, 49: 309-13.

Twycross RG (1974) Diamorphine and cocaine elixir BPC 1973. Pharmaceutical Journal 212(5755): 153 and 159.

Twycross RG, Fry DE and Wills PD (1974) The alimentary absorption of diamorphine and morphine in man as indicated by urinary excretion studies. British Journal of Clinical Pharmacology, 1: 491-4.

Twycross RG (1977a) Choice of strong analgesic in terminal cancer: diamorphine or morphine? Pain, 3: 93-104.

Twycross RG (1977b) Value of cocaine in opiate-containing elixirs. British Medical Journal, Letter, 2: 1348.
Twycross RG (1979) The Brompton Cocktail. In JJ Bonica and V Ventafridda eds Advances in Pain Research and Therapy, Vol 2. New York: Raven Press, 291-300.

Twycross RG (1996) Hospice History Project interview with David Clark.




6 Responses to The Brompton Cocktail: 19th century origins to 20th century demise – by David Clark Comments (RSS) Comments (RSS)

  1. The Brompton Cocktail was still in use in the mid 90 s in Hull , East Yorkshire . I remember patients being admitted to the hospice around 1995 with bottles of it ! Some GP s swore by it !

  2. Back in the 1960’s and 70’s I recall dispensing Brompton Cocktail. Often it contained heroin, Cocaine, 90% alcohol, honey and orange juice.

  3. In 1970’s to 1990’s I prepared Brompton cocktail with honey and brandy as ingredients with heroine and cocaine. On one occasion I had to ring the GP to explain that he could not simply write Brompton cocktail as though it was a Bloody Mary but had to write all the details of the ingredients he was ordering

Leave a Reply

Your email address will not be published. Required fields are marked *