Some conscious patients on the ICU request withdrawal of life-sustaining treatment in order to die. In The Netherlands two scenarios can be offered: (1) deep palliative sedation through administration of barbiturates or benzodiazepines before withdrawal of treatment, or (2) we can offer deliberate termination of life (euthanasia) before termination of treatment. The aim of palliative sedation is to prevent suffering by lowering consciousness as a means to achieve this. In this context, it is important to understand that withdrawal of treatment, consequently allowing the patient to die, after intentional lowering of consciousness is not equivalent with euthanasia or deliberate termination of life [1, 2]. Euthanasia is the intentional termination of life of an adult patient on his or her request. This presupposes voluntariness and a deliberate act, and excludes every form of intentional non-voluntary termination of life. Usually the patient is sedated to unconsciousness and then given a barbiturate and muscle-paralysing agent. In the Termination of Life on Request and Assisted Suicide Act, the requirements of due care are described. Deliberate termination of life is illegal in most countries in the world.

Recently, two of our conscious ICU patients asked for discontinuation of life-sustaining measures. We want to emphasise the central place for the autonomous choice of the patient and the care of the patient and his or her benefits and interests.

Patient A

A ventilator- and inotropes-dependent 61-year-old patient with a history of failed mitral valve replacement and intra-operative total right ventricular myocardial infarction on our ICU requested withdrawal of life-sustaining treatment and if necessary to deliberately terminate her life. In a series of extensive deliberations with the patient and after consulting an independent physician, we agreed to offer her, according to her will, deep palliative sedation followed by withdrawal of life-sustaining treatment. In the presence of her relatives, intravenous administration of midazolam (starting dose 5 mg, thereafter 10 mg/h) and fentanyl (100 μg starting dose, with 50 μg/h to follow) was started. Within 15 min she was unconscious. Subsequently, the inotropes, other life-sustaining medication and mechanical ventilation were withdrawn. She died quietly from heart failure and circulatory collapse.

Patient B

Patient B, a 58-year-old building constructor, was admitted to the general ICU with a traumatic spinal cord injury at the C6–C7 level. In the following weeks, we were unable to wean the patient off the ventilator. Furthermore, he showed no improvements of his spinal cord injury and remained tetraplegic. Several weeks after admission, he repeatedly expressed a clear wish to withdraw treatment and asked the ICU team to take him off the mechanical ventilator. Because of the anticipated anguish, he further wished to be deeply sedated beforehand. After consulting two independent physicians, the nursing staff and a clinical ethicist, the patient was granted his wish. A combination of intravenously administered midazolam (starting with a dose of 5 mg/h, thereafter 10 mg/h) and low dose fentanyl (100 μg starting dose, with 50 μg/h to follow) was started, and after he entered a deep sleep, mechanical ventilation was withdrawn. Shortly thereafter, he died peacefully, surrounded by his family.


Approximately 80% of patients who die in an ICU in The Netherlands die after organ function supportive care or when organ function replacement therapy is withdrawn (unpublished data, Erasmus MC, Department of Intensive Care). Those surviving for a number of hours after discontinuation of treatment can develop symptoms such as dyspnoea-associated respiratory distress, terminal restlessness and ‘death rattle’. The majority of terminal patients in the ICU are however unconscious due to the severity of their disease or multiple organ failure, and are not or only superficially aware of the distressing symptoms. The burden of terminal distress in such cases rests predominantly with the relatives [3]. Most symptoms, however, can be successfully prevented or treated [4]. A national guideline in The Netherlands on withdrawal of futile intensive care measures focuses on anticipation of distressing symptoms. Even in The Netherlands, deliberate termination of life of ICU patients is extremely rare [1].

Palliative sedation is initiated in the ICU in the terminal phase of life to prevent or relieve aggravating symptoms such as terminal restlessness, delirium or anxiety. Sedatives are administered aiming to induce a deep unconscious sleep in anticipation of death. In principle, the National Guidelines for Palliative Sedation of the Dutch Royal Society of Medicine (KNMG) [5] are adhered to, based on the understanding that in ICU practice, only midazolam and propofol are used for sedation. Morphine is not an effective sedative and is not used in the ICU as such [6]. In the ICU, the transition from sedation with mechanical ventilation to palliative sedation is usually imperceptible. As a rule, the sedative regime the patient received while being treated with mechanical ventilation will be continued in the palliative phase until death. The dose required in ICU patients can be significantly higher than mentioned in the recommendations in the guidelines. If the patient is adequately sedated before withdrawal of treatment, then this dose is adhered to. It serves no purpose to adjust an adequate sedative dose either to a lower level or to increase it to abide by the guidelines-recommended starting dose. After publication of the guidelines in 2005, palliative sedation was recognised by the Public Prosecutor to be normal medical practice. There is, however, a directive to prosecute and bring before a court anyone who performs an act not stated as such in the professional guidelines and to treat this as a life-terminating act without the consent of the patient. In the ICU, we deviate from the precise statements in the guidelines, but with good reason. During the documented follow-up of the use of sedatives in patients in the ICU, it is important to judge whether one has acted according to the professional standards. Comparing palliative sedation in the ICU with other non-ICU settings is not realistic.

Crucial in the decision-making process in the described cases is respect for the autonomy of the patient. Based on the law, Dutch caregivers have to respect the wishes of the patient if these are understandable and within the accepted possibilities of medical care and also if the patient is judged to be competent to make these judgments. In the illustrated cases, caregivers had no reasons to doubt the cognitive functioning and competency of the patients.

One cannot always predict with certainty whether or not aggravating symptoms will arise after stopping treatment in the ICU. Patients who are awake may prefer to be unconscious during dying with failing organ functions after withdrawal of therapy. It is our moral duty as caregivers in the ICU to acknowledge this request, and applying palliative sedation prior to withdrawing active treatment offers a righteous means [4]. But with this, we digress from the Dutch Guidelines, which state that administering sedatives to patients for the sole reason of the patient’s wish to be free of suspected suffering at the end of life, is not permitted. There must be objective proof that there is unbearable suffering, which only can be treated by lowering consciousness of the patient.

Many published ethical analyses on palliative sedation use the doctrine of double effect (DDE), mentioning that palliative sedation may hasten or induce death, and is therefore morally questionable. When evaluating an action, the DDE distinguishes between intended effects and the consequences that are foreseen but unintended. As long as intentions are good, it is permissible to perform actions with foreseen consequences that would be wrong to intend. In this line of thought, deliberately causing death is morally wrong, even if desired by a competent patient whose suffering cannot be relieved in another way. If the patient dies unintentionally as a consequence of another ethically justified intervention (e.g., withdrawal of futile treatment), the action is morally acceptable. The unintended but foreseen effect must also be proportional to the intended good effects. The DDE is often mentioned in justifying the use of opiates and sedatives at the end of life to relieve suffering. There is a growing body of evidence showing that administration of normal doses of sedatives and opiates at the end of life does not shorten life, but prolongs it [711], and only scarce anecdotal data mention that opioids can hasten death [12], making the DDE in end-of-life care a myth [13]. It is concluded that “Clinical studies and decades of experience by experts in pain management and palliative care have shown that the double effect of pain medication has little basis in medical fact;…the myth of the double effect of pain medication, directly contributing to the maltreatment of suffering at the end of life. It is ironic that an ethical principle that is used to justify adequate opioid analgesics contributes to the maltreatment of pain” [13]. The same could be said about palliative sedation with barbiturates and benzodiazepines. DDE in end-of-life care is nothing more than a myth, and using it to label palliative sedation and administration of opiates at the end of life questionable and only defendable from the perspective of the DDE, unsound.

When a competent patient on the ICU asks for termination of treatment, we should seriously take this request into deliberation. The principle of respecting autonomy is a strong guidance in bioethics.

The American Academy of Neurology published a position paper concerning the care of conscious, competent patients with profound paralysis [14, 15]. They concluded that such patients have the right to make health care decisions about their own lives, including acceptance or refusal of life-sustaining treatment. Once patients have decided to forego life-sustaining treatment, physicians have an ethical obligation to minimise their subsequent suffering. This is particularly true of profoundly paralysed patients, because cognition and sensation may be intact, and they are capable of great suffering.


In competent ICU patients in The Netherlands who are terminally dependent on life-sustaining measures, such as mechanical ventilation and administration of vasoactive agents, withdrawal of these measures is ethically good care when they ask for withdrawal. This is not to be seen as deliberate termination of life, even if they die after the withdrawal of life-sustaining measures. Administration of sedatives and analgetics before withdrawal of treatment is normal palliative care. The DDE is not applicable in this situation.