Pain is a major symptom in 75 per cent of patients with advanced cancer. A sense of hopelessness and the fear of impending death often adds to the total suffering of the patient. It is necessary, therefore, to think in terms of ‘total pain’. This takes cognizance of psychological, social, and spiritual factors, in addition to the underlying physical component.

A 55-year-old man with recently diagnosed cancer of the oesophagus was still in pain despite receiving slow-release morphine tablets, 6000mg (100mg × 60) twice a day. Following admission to a hospice, he became pain-free on morphine 30mg twice a day and diazepam 10mg nocte. He was able to return home, and converted the spare bedroom into a workshop where he spent many happy hours creatively and productively. The key to success was listening, explaining, and setting positive rehabilitation goals.

Pain in cancer may be caused by the cancer itself, related to the cancer or debility (e.g. muscle spasm, constipation, bedsores), related to cancer treatment (e.g. chronic postoperative scar pain, chemotherapy mucositis), or caused by a concurrent disorder (e.g., spondylosis, arthritis).

A prospective survey of 100 cancer patients with pain revealed a total of 303 anatomically distinct pains, an average of three per patient. Eighty patients had more than one pain; 34 had four or more (Fig. 1) 2791. About two-thirds of the pains were caused by the cancer itself (Table 1) 697, but cancer was the sole cause of pain in only 41 patients. The pains experienced by nine patients were all non-malignant in origin. Most other surveys have concentrated on the patient's worst pain, obscuring the fact that the majority of patients suffer multiple pains.

Doctors caring for cancer patients need to acquire the skill of determining the cause of pain on the basis of diagnostic probabilities. Invasive investigations are increasingly contraindicated as patients move closer to death. A comprehensive list of possible causes of pain is unnecessary: it would be too long to be helpful. What is required is an informed imagination. For example, an awareness of the common muscle pain syndromes prevents many erroneous conclusions. Some knowledge of the patterns of metastatic spread is also helpful ( Figs 2 2792,2793, 3). Pain that is wrongly assumed to be cancerous in origin tends to be invested with all the negative implications of cancer pain, which makes the pain much worse.

A detailed history of pain medication is part of the assessment. It includes not only a list of drugs but also dose used, interval between doses, route of administration, whether taken regularly or as needed, patient's view on drug efficacy, adverse effects, duration of use of drug, and reason for discontinuation.

Because not all pains respond equally to analgesics, information about the benefits of specific medication (or lack of it) often helps to pinpoint the underlying mechanism. From a therapeutic point of view, pain in cancer patients may be divided into three categories: opioid responsive pains, opioid semi-responsive pains (pain relieved by the concurrent use of an opioid and an adjuvant drug), and opioid resistant pains (pain that is not relieved by opioids, but is relieved by other drugs). This classification is important: opioids are of limited value for certain pains, and they may be contraindicated (Table 2) 698. The classification holds true, however, only for opioids given by mouth or by conventional parenteral routes. Opioid resistant pains may well be more responsive to opioid analgesics given by the epidural or intrathecal route.

The use of analgesics and pain management are not synonymous: analgesics are only part of a multimodal approach to treatment (Table 3) 699. However, most cancer patients with pain do need analgesics, often for several months.

Visceral pains come into this category. Six important concepts govern the use of analgesics in the management of opioid responsive pains: ‘by the mouth’; ‘by the clock’; ‘by the ladder’; ‘for the individual’; ‘monitor treatment’; and ‘adjuvant medication’.

Morphine and other strong opioids are effective when administered by mouth. Thus, apart from the last few hours or days of life, few patients require injections to control their pain. Because of reduced bioavailability, however, the oral dose is twice the subcutaneous dose or three times the intravenous dose. Patients with intractable vomiting as well as pain need parenteral antiemetic and analgesic medication. Once vomiting has been controlled it is generally possible to revert to administration of analgesics by the oral route.

Allowing pain to re-emerge before administering the next dose of analgesic causes unnecessary suffering and encourages the development of tolerance. ‘As required’ medication regimens have no place in the treatment of persistent pain (Fig. 4) 2794. A patient with continuous pain, whatever the cause, requires regular preventive therapy. Each dose is given before the effect of the previous one has worn off and, therefore, before the patient may think it necessary.

For codeine and morphine administration every 4 h is optimal. If a strong analgesic other than morphine is used, the doctor must be familiar with its pharmacology. For example, pethidine is generally effective for only 2 to 3 h, though it is commonly given every 4 to 6 h. On the other hand, levorphanol and phenazocine produce satisfactory analgesia when given every 6 h, and buprenorphine and methadone can be given every 8 h. Methadone has a plasma half-life of over 2 days when taken regularly by mouth: cumulation is likely, leading to worsening adverse effects, particularly in the elderly and debilitated.

The archetypal analgesics are aspirin, codeine, and morphine. Other analgesics should be considered as alternatives of fashion or convenience. It is necessary to be familiar with one or two alternatives for patients who cannot tolerate the standard preparations. Aspirin has two alternatives: paracetamol (which has no anti-inflammatory effect) and the non-steroidal anti-inflammatory drugs. For step 2, the author uses dextropropoxyphene in preference to codeine. It causes less constipation and, in the United Kingdom, the compound tablet with paracetamol (coproxamol) has a considerably greater ‘codeine-equivalent’ content than most other weak opioid compound tablets.

The use of morphine is determined by analgesic need and not, for example, by the doctor's estimate of life expectancy—which is often wrong. The correct dose of morphine is that which gives adequate relief for 4 h without unacceptable adverse effects. Maximum or recommended doses, derived mainly from postoperative parenteral single dose studies, are not applicable for patients with cancer pain.

It should be noted that it is pharmacological nonsense to prescribe two weak opioids or two strong opioids simultaneously. It is sometimes justifiable for a patient receiving a strong opioid to be given another weak or strong opioid as a second ‘as required’ analgesic for occasional, troublesome pain. However, if pain occurs despite regular analgesia, patients should be advised to take an extra dose of their regular medication. Short-acting preparations such as pentazocine (weak opioid), pethidine (intermediate), and dextromoramide (strong opioid) should be avoided. An opioid mixed agonist-antagonist (e.g. pentazocine) or a partial agonist (buprenorphine) should not be given with a strong opioid agonist (i.e. morphine). It is unnecessary and may result in unintended antagonism.

Morphine by mouth can be given either as an aqueous solution of morphine sulphate or as tablets. In the United Kingdom, and in several other countries, slow release morphine tablets are available as 10, 30, 60, 100, and 200mg strengths. Most patients changing from a weak opioid initially receive 60mg of morphine a day (aqueous morphine sulphate 10mg every 4 h or slow-release morphine 30mg every 12 h). As the effective dose of oral morphine sulphate ranges from as little as 5mg to more than 1g every 4 h, the top of the ladder is not reached simply by prescribing morphine. Most patients, however, never need more than 100mg every 4 h; in the majority pain is well controlled by doses as small as 10 to 30mg (or slow-release morphine 30 to 100mg twice daily).

All cancer patients receiving analgesics need close supervision to achieve maximum comfort with minimal adverse effects. Treatment review is sometimes necessary within hours; it is certainly required after 1 to 2 days and at the end of the first week. Subsequent assessment varies according to psychological and therapeutic needs. New pains may develop and old pains re-emerge: a fresh complaint of pain demands reassessment, not just a message to increase pain medication, though this may be an important short-term measure.

A laxative is almost always necessary for patients receiving an opioid. Experience at many centres strongly suggests that a combination of a peristaltic stimulant and a surface-wetting agent achieves the best results, for example, standardized senna (up to 75mg a day) and docusate (up to 600mg a day). Combination preparations may reduce the number of tablets or capsules the patient has to swallow (e.g. codanthrusate up to 6 capsules a day). More than one-third of in patients with terminal cancer need rectal measures (suppositories, enemas, or manual evacuations) in addition to oral laxatives.

Haloperidol 1 to 1.5mg immediately and at night is the treatment of choice for opioid-induced vomiting. In some patients receiving morphine, haloperidol is ineffective because of morphine-induced delayed gastric emptying. Thus, if a patient does not respond to haloperidol and the pattern of vomiting is suggestive of gastric stasis, a gastrokinetic antiemetic (metoclopramide or domperidone 10mg up to every 4 h) should be substituted.

If a patient is very anxious, an anxiolytic should be prescribed (e.g. diazepam 5mg immediately and 5 - 10mg at night). If a patient remains depressed after 2 or more weeks of improved pain relief and psychological support, an antidepressant should be tried. Many patients benefit from the use of a night sedative (e.g. temazepam, chloral hydrate).

Functional gastrointestinal pains, although opioid responsive, are best treated by other means. Colic associated with constipation, for example, should be treated by relieving the constipation, as should rectal spasm caused by faecal impaction. Constipation can also cause pain in the right iliac fossa. In this circumstance, the descending colon (and sometimes the transverse colon) is easily palpable because it is full of hard faeces. The caecum feels distended and is tender on palpation. The pain is that of gaseous caecal distension. Identical caecal symptoms and signs may also be seen in patients with obstruction of the colon. Careful history taking and assessment normally indicates which is the most likely of the two diagnoses. Sometimes the intensity of the discomfort may warrant administration of a weak opioid preparation (e.g. coproxamol) for several days while the constipation is being corrected.

In patients with moderately severe colic associated with partial intestinal obstruction, an antispasmodic may be adequate (e.g. mebeverine 135 - 270mg four times daily). Occasional severe colic is sometimes best treated with sublingual hyoscine hydrobromide 0.3mg as needed, which acts within minutes of administration. Hyoscine butylbromide (80 - 160 mg) by continuous subcutaneous infusion is sometimes necessary.

‘Squashed stomach syndrome’ is common in patients with advanced cancer. In this syndrome, epigastric pain is caused by relative gastric distension. This often occurs in patients with a grossly enlarged liver, whether or not there is any associated gastric abnormality. It is important to recognize the postprandial discomfort for what it is because explanation to the patient is crucial in management. Some patients also experience retrosternal pain due to acid reflux. The aim of treatment is to prevent distension. This requires a combination of dietary measures (as after partial gastrectomy) and drug treatments (an antiflatulent such as dimethicone and possibly a prokinetic agent such as metoclopramide).

Irritable bowel syndrome occurs in about 10 per cent of cancer patients (as in the general population) and requires specific treatment.

Pain caused by bone metastases is often only partially opioid responsive. Best results are obtained with a combination of a non-steroidal anti-inflammatory agent and morphine. Many osseous metastases produce or induce the production of a prostaglandin which causes osteolysis and also lowers the peripheral pain threshold by sensitizing the nerve endings. Non-steroidal anti-inflammatory agents inhibit the synthesis of prostaglandins and thereby alleviate pain.

Response to prostaglandin inhibitors is variable. Even so, a non-steroidal anti-inflammatory drug should be administered routinely for the relief of bone pain, either alone or in combination with an opioid. Generally, one of the newer twice-daily agents should be prescribed (flurbiprofen, diflunisal, or naproxen).

Pain caused by nerve compression is often not controlled by morphine alone. In this situation a corticosteroid such as dexamethasone 4 to 8mg daily should be prescribed. A marked improvement is often seen within 48 h. If the nerve compression relates to an identifiable bone metastasis or soft tissue mass, radiotherapy should be considered: this normally allows the dose of morphine and dexamethasone to be reduced. A neurolytic or neuroablative procedure should be considered in patients who do not respond adequately to combined morphine and dexamethasone. Such procedures are necessary in less than 10 per cent of all cancer pain patients.

A corticoid should be considered as a ‘coanalgesic’ whenever there is a large tumour mass within a relatively confined space (Table 4) 700. A tumour is often surrounded by inflamed oedematous tissue, and pressure on neighbouring veins and lymphatics may lead to further local or regional swelling: by reducing the inflammation, corticoids thereby reduce the total tumour mass.

Corticosteroids reduce the production of prostaglandins by inhibiting phospholipase activity, and thereby preventing the formation of arachidonic acid, the precursor substance of prostaglandins, from cell membrane phospholipids. There is, therefore, a theoretical case for using both corticosteroids and non-steroidal anti-inflammatory drugs in the management of certain pains, such as those due to bone metastases. However, more tablets may lead to poorer compliance and more adverse effects. Sometimes, however, maximal relief is obtained only by concurrent use of an opioid, a non-steroidal anti-inflammatory agent, and a corticosteroid, for example when pelvic malignancy invades bone and compresses nerves.

A severe muscle cramp is a universal experience. The pain is extremely intense for a short time: it is normally coped with partly by the knowledge that, ‘it is only cramp’. Muscle pain in cancer patients, unless recognized for what it is, is likely to be thought to be due to cancer. Such a response will inevitably magnify the pain and perpetuate it.

Muscle spasm pain secondary to underlying bone pain and/or skeletal deformity is common in cancer patients. Myofascial trigger point-related pains also occur. The correct approach is explanation, physical therapies (local heat and massage), diazepam and relaxation therapy, and injection of the trigger points with local anaesthestic and a corticosteroid (e.g. 0.5 per cent bupivacaine and depot methylprednisolone). However severe, morphine is not indicated for relief of muscle spasm and trigger point pains.

Neural injury pain is usually peripheral, resulting from infiltration of a nerve in cancer. In this circumstance, the pain is dermatomal in distribution, and is usually superficial and burning in character. There is often associated allodynia, i.e. light touch and other non-noxious stimuli cause pain. There may also be stabbing pain. Neural injury pains do not always respond to opioids. They are helped, however, by an antidepressant and/or an anticonvulsant, possibly used in combination with morphine—depending on circumstances and response.

Very occasionally a cancer patient presents with an opioid resistant pain which is dependent on the integrity of the regional sympathetic nerves. The essential features are pain (often burning), a sensory disorder, and relief following a regional sympathetic nerve block.

It is necessary to differentiate between neural injury burning pain and sympathetically-mediated burning pain. This may be difficult as features are not constant and some are common to both conditions (Table 5) 701. However, because sympathetic nerves travel with the arterial system, the distribution of sympathetically-mediated pain is vasotopographic. There may also be radiographic evidence of osteoporosis and of hot spots on isotope bone scans; these may be mistaken for osteolytic metastases,

If sympathetically-mediated pain is suspected in a cancer patient, a diagnostic sympathetic block should be performed with local anaesthetic. This not only serves to confirm the diagnosis but may also give relief that outlasts the duration of the local anaesthetic effect. A neurolytic block should be considered at a later date for lower limb pain if the pain returns.

Satisfactory pain relief can be achieved in the majority of cancer patients. The patient's pain(s) must, however, be assessed rapidly and accurately. The doctor must appreciate that pain is a psychosomatic experience, and understand the pathological processes which give rise to pain, the neurological anatomy, and the phenomenon of referred pain. He must be aware that one-third of pains experienced by patients with advanced cancer are not caused directly by the cancer, and be able to recognize pain caused by muscle spasm. Differentiation between opioid responsive and opioid resistant pains is important to allow optimum treatment, as is familiarity with the use of antidepressants and anticonvulsants in neuropathic pain, and knowledge of the benefits of corticosteroids with certain pains. Patients prescribed analgesics or other drugs must be closely supervised to assess benefit, monitor adverse effects, and to modify treatment if necessary.

Arner S, Meyerson BA. Lack of analgesic effect of opioids on neuropathic and idiopathic forms of pain. Pain 1988; 33:11 - 23.

Churcher MD, Ingall JRF. Sympathetic dependent pain. Pain Clinics 1987; 1:217 - 18.

Ferreira SH. Prostaglandins, aspirin-like drugs and analgesics. Nature 1972; 240:200 - 3.

Galasko CSB. The development of skeletal metastases. In: Weiss L, Gilbert HA, eds. Bone Metastasis Boston: GK Hall, 1981:83 - 113.

Hanks GW, Hoskin PJ, Aherne GW, Turner P. Explanation for potency of repeated oral doses of morphine? Lancet 1987; ii:723 - 4.

McQuay J. Pharmacological treatment of neuralgic and neuropathic pain. Cancer Surv 1988; 7:141 - 59.

Meyerson BA. Neuropathic pain: an overview. In: Lipton S, Tunics E, eds. Advances in Pain Research and Therapy. Volume 13. New York: Raven Press, 1990:193 - 9.

Ottesen S, Minton M, Twycross RG. The use of epidural morphine at a palliative care centre. Palliative Medicine 1990; 4:117 - 22.

Raftery H. The management of postherpetic pain using sodium valproate and amitriptyline. Ir Med J 1979; 72:399 - 401.

Read NW. Irritable Bowel Syndrome. London: Grune and Stratton, 1985.

Saunders C, Baines M. Living with dying. The Management of Terminal Disease. Oxford: Oxford University Press, 1983.

Shutt LE. Treatment of autonomically mediated pain. Curr Opin Anaesthesiol 1989; 2:627 - 31.

Takeda F. The management of cancer pain in Japan. In: Twycross RG, ed. The Edinburgh Symposium on Pain Control and Medical Education. London: Royal Society of Medicine, 1989:17 - 21.

Tasker RR. The problem of deafferentation pain in the management of the patient with cancer. J Palliative Care 1987; 2:8 - 12.

Twycross RG, Fairfield S. Pain in far-advanced cancer. Pain 1982; 14:303 - 10.

Twycross RG, Lack SA. Symptom Control in Far-Advanced Cancer: Pain Relief. London: Pitman, 1983.

Twycross RG, Lack SA. Alimentary Symptoms in Far-Advanced Cancer. Edinburgh: Churchill Livingstone, 1986.

Twycross RG, Lack SA. Therapeutics in Terminal Cancer. 2nd edn Edinburgh: Churchill Livingstone, 1990.

Ventafridda V. Continuing care: a major issue in cancer pain management. Pain 1989; 36:137 - 43.

Ventafridda V, Tamburini M, Caraceni A, De Conno F, Naldi F. A validation study of the W.H.O. method for cancer pain relief. Cancer 1987; 59:851 - 6.

World Health Organization. Cancer pain Relief. Geneva: World Health Organization, 1986.