The majority of dying patients in surgical wards have cancer. The text is written with such patients specifically in mind, though the underlying principles have more general application. When caring for a dying patient, the main goals are relief of pain and other distressing symptoms, and psychological and spiritual care for patients so that they may come to terms with their own death as fully and constructively as possible. For patients who are terminally ill rather than imminently dying, a support system to help patients live as actively and creatively as possible, thereby facilitating autonomy and enhancing self-esteem, is also important. It is also necessary to provide parallel support for the patient's family. Care of the dying patient therefore calls for a comprehensive approach, encompassing physical, psychological, social, and spiritual aspects (Table 1) 691. This requires multiprofessional teamwork.

Medical care is a continuum, ranging from complete cure at one end to symptom control and support at the other. Many types of treatment span the entire spectrum of care, notably radiotherapy and, to a lesser extent, chemotherapy and surgery. It is important, therefore, to keep the therapeutic aim clearly in mind when employing any form of treatment. The key points to bear in mind are the patient's biological prospects, the therapeutic aim of each treatment, the adverse effects of treatment, the benefits of treatment for the patient, and the need not to prescribe a lingering death.

When a person is close to death the primary aim of treatment is comfort. What is appropriate for an acutely ill patient may well be inappropriate in the dying. Nasogastric tubes, intravenous infusions, antibiotics, cardiac resuscitation, and artificial respiration are primarily supportive measures to assist a patient with acute or acute-on-chronic illnesses through the initial period towards recovery of health. Such measures are inappropriate in patients who are close to death and for whom there is no expectation of a return to health if they simply prolong the process and distress of dying.

Although unexpected improvement is sometimes seen, there are many occasions when it is appropriate to ‘give death a chance’. All patients must die eventually; ultimately nature will take its course. The art of medicine is to decide when life sustenance is essentially futile and, therefore, when to allow death to occur without further impediment. Antibiotics are usually appropriate for the patient with advanced cancer who develops a chest infection while still relatively active and independent. Pneumonia should, however, still be ‘the old man's friend’. If the burden of living has become considerable, it may be more appropriate not to use antibiotics. When it is difficult to make a decision, the ‘two day rule’ should be invoked: if after 2 or 3 days of straightforward symptom relief the patient is clearly holding his own, prescribe an antibiotic. If the patient is clearly much worse, do not.

As patients become bedbound, it is sensible to phase out the less important medications. Multivitamins, iron, and potassium supplements are all medium or long-term supportive treatments; they are irrelevant when the patient is close to death. Antihypertensives can often be phased out as weight loss and/or other medication alter the patient's need; for example, morphine reduces the cardiac preload. Continuation of pre-existing diuretic therapy may cause postural hypotension unless the dose is reduced or stopped in the face of a poor fluid intake or repeated vomiting. When death seems close, giving the patient ‘permission’ not to eat and discouraging the family from force-feeding may well be appropriate.

A wide variety of symptoms is experienced by patients with advanced cancer, but no symptom occurs in every patient. Even pain is experienced by only about three-quarters of patients (Section 44.2) 383. Not all symptoms are caused by the cancer. They may also derive from treatment (past or present) or a concurrent second disorder; some symptoms are caused by multiple factors. It is necessary to identify the cause and seek to correct those that are reversible. In this way, although the underlying pathological process remains unchanged, it is generally possible to obtain significant, if not complete, relief.

If pharmacological treatment of a persistent symptom is appropriate, the drug(s) should be prescribed regularly on a prophylactic basis. Although this concept is widely applied in relation to the use of analgesics, its relevance to the use of antiemetics and laxatives is not always appreciated. The use of drugs ‘as needed’ instead of regularly ‘by the clock’ is the cause of much unrelieved distress.

As the needs of patients vary, it is often impossible to predict the optimum dose of, for example, opioids, laxatives, and psychotropic drugs. Adverse effects may also jeopardize patient compliance. Adjustments are therefore necessary, particularly during the first few weeks. This should be anticipated and arrangements made for continuing close supervision.

Although many symptoms respond to a combination of drug and non-drug measures, it is sometimes necessary to compromise in order to avoid unacceptable adverse effects. For example, anticholinergic effects such as dry mouth or visual disturbance may be limiting factors. In patients with intestinal obstruction, it is often better to aim to reduce the incidence of vomiting to once or twice a day rather than to seek absolute control.

Nausea and/or vomiting occurs in about 40 per cent of patients with terminal cancer. It is important to choose the most appropriate antiemetic (Table 2) 692. A patient whose vomiting is a result of intestinal obstruction or raised intracranial pressure does not respond well to a drug which acts principally (e.g. prochlorperazine) or exclusively (e.g. haloperidol) on the chemoreceptor trigger zone in the floor of the fourth ventricle. If the vomiting is related to physical or chemical irritation of the stomach, a prokinetic antiemetic such as metoclopramide or domperidone is the best choice (Table 3) 693. These enhance peristalsis in the upper gastrointestinal tract, speed gastric emptying, and increasing the tone of the oesophageal sphincter.

Antagonists of the 5-hydroxytryptamine type 3 receptor, such as ondansetron and granisetron, are used to control vomiting caused by highly emetogenic chemotherapeutic agents (such as cisplatin, Adriamycin, and cyclophosphamide) and by upper abdominal and upper lumbar spine irradiation. These act both in the upper intestinal tract, where they antagonize the action on the vagal nerve endings of 5-hydroxytryptamine released from enterochromaffin cells, and in the chemoreceptor trigger zone.

When the gastrointestinal tract is obstructed and surgery is contraindicated, an antiemetic such as cyclizine or chlorpromazine should be prescribed. There is some evidence that metoclopramide (a prokinetic antiemetic) may also be helpful. With a high obstruction (gastric outlet or duodenum), metoclopramide may, however, exacerbate rather than reduce vomiting. It is important, however, to reduce gastrointestinal motility in order to prevent the colic which often precedes and possibly triggers obstructive vomiting.

Unfortunately, no two patients with obstruction are alike. It is necessary to consider each patient individually, and proceed with a series of therapeutic trials. If the obstruction is in the pylorus, duodenum, or jejunum, a nasogastric tube may be necessary. This can be aspirated before drinks and meals, and possibly left on free drainage overnight. The symptoms will be greatly relieved and the patient often remains remarkably well for several days before deteriorating rapidly as a result of accumulated fluid and chemical losses.

As the dying patient becomes weaker, he naturally drinks less. This may relate to dysphagia, nausea, anorexia, or lack of energy and interest. A reduced level of consciousness may also be relevant. As a general rule intravenous fluids are not administered to patients dying in palliative care centres. On the other hand, patients dying in a general hospital frequently continue to receive intravenous fluids. This reflects a belief that dehydration is distressing, and therefore unacceptable. Such a view has, however, been disputed. Indeed, in the patient close to death, dehydration may be beneficial. As body fluid decreases, vomiting in patients with gastrointestinal obstruction may be reduced to only once or twice a day, even without specific symptom control measures. In contrast, intravenous fluids will tend to maintain secretions, resulting in more copious and more frequent vomiting. A nasogastric tube is likely to become necessary, whereas it is hardly ever needed if the patient is relatively dehydrated. Moreover, as body fluids reduce, urine output decreases. This results in fewer bed-wetting episodes, and less need for urinal, bedpan, commode, or catheterization.

Perhaps the most discomforting consequence of dehydration is a dry mouth, which may lead to cracking of the mucosa, inflammation, and infection. Dry mouth, however, can be alleviated to a considerable extent by conscientious mouth care. It is often possible to enlist the family's help in this. A bottle with a spray top allows small volumes (1 - 2 ml) to be introduced into the mouth at 20- or 30-min intervals if necessary.

Neuromuscular irritability is the other relatively common effect of marked dehydration. This relates to electrolyte imbalance, and may lead to multifocal myoclonus (muscle twitching) and general restlessness. Diazepam orally or rectally should be given to control these. Alternatively midazolam can be given parenterally, preferably by continuous subcutaneous infusion using a portable syringe driver.

Dehydration should therefore be looked upon as a natural process that helps to relieve a number of symptoms in the last few days of life. Intravenous fluids should be avoided as they tend to exacerbate discomfort in the patient close to death and may also add to the emotional distress of the family.

A confusional state is characterized by ‘clouding of the sensorium’, is usually associated with drowsiness, and manifests as one or more of the following:

poor concentration

impairment of recent memory

disorientation

illusions (often) and paranoid delusions

hallucinations (sometimes)

rambling and incoherent speech

restlessness and noisy aggressive behaviour.

Confusional states vary in degree and intensity. In the dying, most are caused by multiple factors (Table 5) 695. In some patients, symptoms are never more than mild and remain intermittent.

Unfortunately, any sort of strange behaviour or talk in an ill patient tends to be labelled ‘confusion’. Once a patient is deemed confused, doctors, nurses, friends, and family tend to shy away in embarrassment or even in fear. Precise assessment enables the family and carers to continue to respond supportively to someone who happens to be very ill and who is experiencing what is often only a minor disturbance of thought.

Disorientation for time and, to a lesser extent, for place is the norm for the dying. Explanation and reassurance may be all that is necessary. Hallucinations, illusions, nightmares, and vivid daydreams often reflect anxiety about death. Anxiolytic medication, together with continuing psychological support, is helpful when the content suggests this. Blunderbuss therapy with chlorpromazine is a common medical response to this situation: although this may help to suppress hallucinations, it may increase the tendency to misperceive as a result of further clouding of the sensorium (Table 6) 696.

Hope is an expectation greater than zero of achieving a goal. Hope tends to diminish when:

the patient is mentally isolated by a ‘conspiracy of silence’;

it is implied ‘there is nothing more that can be done’;

pain and other symptoms remain unrelieved or ignored;

the patient feels alone or unsupported.

Not destroying hope is frequently used as a reason for not informing a patient of the seriousness of his situation. False optimism, however, is a potent destroyer of hope. On the other hand, an unwise catharsis by the doctor of all that is negative, either to the patient or to the family, may make the doctor feel easier but it can irreversibly destroy hope and result in intractable anxiety and despair. It is necessary to apply two parallel principles: ‘Never lie to a patient’, and ‘Avoid total candour.’ Most people are best served when realism is tinged with optimism. Gentle, sympathetic, and gradual communication of the truth, within the context of continued support and encouragement, almost never destroys hope. In fact, good terminal care restores hope by encouraging realistic goals. Telling the patient what you plan to do about his pain and other symptoms helps to counter the living nightmare of never-ending unendurable pain or other physical distress.

When dying many people take stock of their lives for the first time;

‘I have lived a good life’;

‘I never did anyone any harm’;

‘Why did it happen to me?’

Although only a minority of patients discuss these issues with their doctor, the majority do so with a nurse or social worker, or with their relatives and close friends. Patients are very perceptive, and they are unlikely to embarrass a doctor if they sense that communication at this level will cause him discomfort.

When possible and appropriate, it is right for the doctor to alert the chaplain, priest, or rabbi to the fact that the patient is seriously ill and may appreciate a visit. Although regard for the patient as an individual does not allow the imposition of one's own beliefs, many patients are comforted by the discovery that their doctor has a religious faith.

As the patient becomes increasingly weak he is faced with the fact that death is inevitable and imminent. This is a time when support and companionship are of paramount importance. Although the doctor may feel powerless in the face of rapidly approaching death, the patient is usually more realistic. He knows a miracle cannot be performed, he realizes that his time has come, and, provided he is relatively comfortable, continued medical attentiveness is all he expects. The situation need never become negative and hopeless provided the doctor continues to visit;

quietly indicates that ‘at this stage the important thing is to keep you as comfortable as possible’;

simplifies medication as much as possible;

arranges for medication to be given by suppository or injection when the patient is too weak to swallow;

continues to inform the family of the changing situation;

takes steps to control an agitated confusional state even if it results in greater somnolence;

listens to the nurses.

The care of the family is an integral part of the care of the dying. Family - doctor communication generally needs to be initiated and maintained by the doctor. It is easy to neglect the relatives, who are often reluctant to bother a doctor who they see as busy. There is much to be said, at the time of diagnosis and later, for joint interviews with the patient and spouse. The doctor should also make an opportunity to see both the patient and the close relatives on their own. Further separate or joint interviews can be arranged as necessary.

It is not generally necessary or wise to tell the family the whole truth (as you see it) at one time. If the family and patient are too far ‘out of step’ in relation to knowledge about the diagnosis and prognosis, it can create a barrier between the two. They commonly ask for the patient not to be told. This should be seen as the initial shock reaction and not as an excuse for saying nothing to the patient. If the family and patient are to be mutually supportive, the relatives must be helped to move forward from this initial reaction to a position of greater openness and trust. It is important to remember that the family cannot forbid the doctor to discuss diagnosis and prognosis with a patient.

Admission to hospital is often seen as a defeat by the family. It is necessary to emphasize that you are surprised that they managed to cope for so long and that now, with the need for day and night care, it is impossible for one (or two) people to continue to cope alone without a break. Frequent visits should be encouraged. Separation anxiety may be reduced by encouraging them to help in the care of the patient by adjusting the pillows, refilling a water jug, helping with a blanket bath, and assisting at mealtimes. Some relatives need to be taught how to visit, to behave as they would at home. For example, sit and read a book or newspaper, knit, watch television together. It should be emphasized that they do not have to keep up a tiring patter of conversation about trivia. If necessary, overnight accommodation should be arranged to enable a close relative to sleep nearby or with the patient in a single room.

Baines M, Oliver, DJ, Carter, RL. Medical management of intestinal obstruction in patients with advanced malignant disease. A clinical and pathological study. Lancet 1985; ii: 990 - 3.

Editorial. Terminal dehydration. Lancet 1986; i: 306.

Gralla RJ. Metoclopramide: a review of antiemetic trials. Drugs 1983; 25: 63 - 73.

Isbister WH, Elder P, Symons L. Non-operative management of malignant intestinal obstruction. J R Coll Surg Edinb 1990; 35: 369 - 72.

Marty M. Ondansetron in the prophylaxis of acute cisplatin-induced nausea and vomiting. Eur J Clin Oncol 1989; 25: S41 - S45.

Peroutka SJ, Snyder SH. Antiemetics: neurotransmitter receptor binding predicts therapeutic actions. Lancet 1982; i: 658 - 9.

Priestman TJ. Clinical studies with ondansetron in the control of radiation-induced emesis. Eur J Clin Oncol 1989; 25: S29 - S33.

Stedeford A. Facing death: patients, families and professionals. London: Heinemann, 1984.

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

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

Tyers MB, Bunce KT, Humphrey PPA. Pharmacological and anti-emetic properties of ondansetron. Eur J Clin Oncol 1989; 25: S15 - S19.

World Health Organization. Palliative Cancer Care. Copenhagen: WHO Regional Office for Europe, 1989.