August 2025
INTRODUCTION —
Withdrawal of ventilatory support is common in the intensive care unit (ICU).
This topic reviews our approach to ventilator withdrawal in patients whose goals of care are focused on palliation of symptoms as opposed to prolonging life.
Other palliative care issues that arise in the ICU are discussed separately:
●(See "Communication in the ICU: Holding a meeting with families and caregivers of adult patients".)
●(See "Palliative care: Issues in the intensive care unit in adults".)
●(See "Informed procedural consent in the intensive care unit".)
The approach outlined in this topic is generally consistent with published guidelines issued by several organizations including the Canadian Critical Care Society (CCCS), the European Society of Intensive Care Medicine (ESICM), and the American Thoracic Society (ATS) [1-7]. (See 'Society guideline links' below.)
GOALS OF CARE DISCUSSIONS
General principles — Discussions about patients' goals of care and preferences around invasive ventilatory support ideally begin well before patients develop impending respiratory failure. Details regarding discussing goals of care, holding family meetings in the intensive care unit, and resolving conflict about treatment near the end of life are described separately:
●(See "Discussing goals of care".)
Ethical issues specific to withdrawing ventilatory care — Specific to patients in whom withdrawing ventilatory support is being considered are the following ethical principles:
●Withdrawing mechanical ventilation is not abandonment – Ventilator withdrawal is not abandonment because the clinician continues a therapeutic relationship with the patient, including intensive symptom management and psychosocial support.
When discussing the withdrawal of mechanical ventilation with patients, families, and other clinicians, we avoid the phrase "withdrawal of care" because it is factually inaccurate and may cause the patient/family to feel abandoned. Instead, we use more precise phrases (eg, "withdrawal of life support" or "withdrawal of the ventilator").
●When aligned with the patient's wishes, withdrawal of mechanical ventilation is ethically acceptable – Mechanical ventilation withdrawal may be mistakenly perceived as a nonbeneficent act on the view that the only way to benefit a patient is to extend life and forestall death. However, the ethical principle of beneficence requires clinicians to promote a patient's well-being, as defined by the patient. For some critically ill patients, ventilator withdrawal best advances the patient's interests by allowing them to die in a way that is consistent with their values and preferences [8-10]. (See "Ethical issues in palliative care".)
●Both withholding and withdrawing ventilatory support are ethically permitted – In multiple countries, including the United States and the United Kingdom, there is no ethically meaningful distinction between withholding and withdrawing life-sustaining treatment. Both are permitted when doing so reflects the informed decision of a patient or their surrogate decision-maker [11]. Withholding ventilatory support is sometimes seen as emotionally easier to accept compared with withdrawing support, probably because there is a perception of less involvement in the patient's death.
Notably, if withholding treatment was permitted but withdrawing treatment was not, patients would not be permitted to request a time-limited trial of intensive care because, once started, ventilatory support could not be withdrawn. Consequently, patients would be required to make an ''all-or-nothing'' decision in the early stage of an illness to either refuse ventilatory support or to authorize its indefinite use. This is ethically problematic because it would pressure patients who might desire a time-limited trial of ventilatory support to forego it out of concern they would be kept alive on a ventilator indefinitely, which they might view as a ''state worse than death'' [8].
●It is ethically acceptable to provide treatment to mitigate distressing symptoms in patients near the end of life, even if doing so may hasten the patient's death – In patients near the end of life, clinicians have an ethical obligation to treat pain and other distressing symptoms, even if doing so might hasten the patient's death. This principle of "double effect" (ie, symptom palliation and hastening of death) is one justification for administering such treatments, which has been incorporated into rulings by the United States Supreme Court [12].
The principle of double effect and how it applies to patients receiving opioids at the end of life is discussed in more detail elsewhere. (See "Ethical considerations in effective pain management at the end of life", section on 'Opioid use' and "Ethical considerations in effective pain management at the end of life", section on 'Principle of double effect'.)
PRACTICAL ASPECTS OF WITHDRAWING MECHANICAL VENTILATION
Protocolized and multidisciplinary approach — Institutions should develop protocols to guide palliative mechanical ventilation withdrawal. The protocol outlined in the sections below is one such example (table 1).
Limited evidence supports a protocolized approach:
●In a stepped wedge cluster randomized trial of 168 patients undergoing palliative withdrawal of mechanical ventilation, a patient-centered algorithmic approach reduced patients' respiratory distress using lower total amounts of opioids and benzodiazepines compared with one-step ventilator withdrawal [13]. The algorithmic approach involved rapid reduction in ventilatory support (mean time 18 min +/- 27 min) paired with a nurse-driven algorithm that used the Respiratory Distress Observation Score to monitor and treat respiratory distress.
●In a prospective trial of 875 family/caregivers/surrogates of patients undergoing ventilator withdrawal, a three-step, physician-driven, nurse-aided support strategy was compared with standard care [14]. The first step was an end-of-life conversation, the second step was the physical presence of health care personnel in the room during the dying process, and the third step was meeting with family/caregivers/surrogates to express condolences after death. At six months, the three-step program reduced the number of family/caregivers/surrogates with prolonged grief (PG) symptoms (21 versus 15 percent) and the median PG-13 score (19 versus 21).
We use a multidisciplinary team approach that includes physicians, nurses, and respiratory therapists. We tailor aspects of the withdrawal protocol to reflect the patient's condition and relevant knowledge of their end-of-life preferences. For example, some patients may place a very high priority on maintaining as much mental clarity as possible during the process of ventilator withdrawal, even if doing so may lead to more physical symptoms, such as breathlessness [15]. Other patients value maximal symptom control and accept the risk of decreased consciousness from sedation and analgesia to accomplish this.
Simulation laboratories are available to train clinicians in palliative withdrawal of mechanical ventilation [16].
Preparing the family or caregivers — Once the decision has been made to withdraw mechanical ventilation, we prepare the family/caregivers/surrogates by explaining each step in advance:
●We offer families/caregivers/surrogates the opportunity to gather at the bedside in advance of the process. We reassure them that withdrawing mechanical ventilation does not have to be immediate.
●We encourage families/caregivers/surrogates to spend time with their loved one before withdrawal of mechanical ventilation to say goodbye or perform rituals or prayers.
●We offer families/caregivers/surrogates the opportunity to meet with hospital chaplains and social workers for spiritual and psychological support.
●We explain to family/caregivers/surrogates the anticipated changes in the respiratory pattern (eg, Cheyne-Stokes breathing) or other involuntary movements (eg, clonus, sighs, rattling) that may occur after ventilator withdrawal. We attempt to estimate if patients may breathe for prolonged periods (eg, days) or pass away quickly (eg, minutes to hours), and we explain that sometimes our estimates are inaccurate. Predictors of time to asystole are discussed below. (See 'Predicting time to asystole' below.)
●After spending time with their loved one, the family/caregivers/surrogates usually indicate to the team their readiness to begin the process. In some cases, the family/caregivers/surrogates are present throughout the ventilator withdrawal process (see 'Gradual weaning' below), but many clinicians ask family members and caregivers to step outside the room briefly while the patient undergoes extubation. When the family/caregivers/surrogates request to remain in the room, we inform them about what they will see. (See 'Procedure of ventilatory support withdrawal' below.)
Preparing the patient — To adequately prepare the patient we perform the following [17]:
●We discontinue neuromuscular blocking (NMB) agents well in advance and in accordance with their half-life (2 to 18 hours) (table 2). Neuromuscular blockade masks signs of discomfort and, therefore, increases the risk that patients will die with inadequately treated symptoms. We may use nerve stimulation devices to ensure no residual NMB agent is present. (See "Neuromuscular blocking agents in critically ill patients: Use, agent selection, administration, and adverse effects", section on 'Monitoring'.)
In rare cases, the NMB agent may be reversed pharmacologically. (See "Clinical use of neuromuscular blocking agents in anesthesia", section on 'Reversal of neuromuscular block'.)
In the very rare case in which it is felt that the return of neuromuscular function would be unduly burdensome to the patient or would cause intractable distress from neuromuscular abnormalities, we advise expert consultation (eg, palliative care or ethics consultants) [1].
●We ensure do-not-resuscitate/intubate (DNR/I) orders have been documented.
●We ensure that intravenous access is secured and suction equipment is ready.
●We ensure that rapid-acting intravenous analgesics and sedatives are at the bedside to quickly alleviate patient distress during weaning and cessation of ventilator support (eg, fentanyl, lorazepam). We anticipate which continuous infusions (if any) will be needed and prepare accordingly (eg, morphine, fentanyl, midazolam). Many patients are often receiving and are comfortable on symptom-targeted analgesic/sedative infusions before ventilator withdrawal. We may use these to help control symptoms (bolus and increased infusion doses). (See "Palliative care: The last hours and days of life", section on 'Management of specific symptoms'.)
●For patients with excessive secretions, we may administer intravenous or subcutaneous glycopyrrolate (glycopyrronium) 0.1 to 0.2 mg every four to eight hours as needed prior to withdrawal for secretion control after withdrawal. Alternatives include a retroauricular 1 mg scopolamine patch (release 1 mg over three days), 1% atropine (eye drops) one to two drops sublingually every six hours as needed, or scopolamine butylbromide 20 mg intravenously or subcutaneously every four to six hours as needed. These agents may be more comfortable for patients as compared with suctioning, although comparative data are lacking.
●We remove soft ties, mitts, and other devices that hinder contact between the patient and their family.
●We typically turn off radios and televisions unless the family/caregivers/surrogates desire them. We also discontinue room alarms and telemetry monitoring screens and provide adequate space and seating for visitors inside or outside the room. We also provide a "comfort cart" with refreshments and snacks outside the room.
Procedure of ventilatory support withdrawal
Choosing among the approaches — There are two main approaches to palliative withdrawal of mechanical ventilation:
●Gradual weaning – This involves a stepwise reduction in ventilatory support and oxygen delivery, with patient assessments at each step paired with the administration of medications for comfort as needed. It is followed by extubation when the patient is comfortable on minimal ventilatory support. Details of the procedure are provided below. (See 'Gradual weaning' below.)
●Immediate extubation – This involves a single step of extubation, without any preceding reductions in the ventilatory support or oxygen delivery. Details of the procedure are provided below. (See 'One-step weaning (immediate extubation)' below.)
In most patients, we prefer gradual weaning because it is more likely to minimize patient discomfort, especially for conscious patients receiving high levels of ventilatory support. Immediate extubation is appropriate for patients with severe neurologic injuries who have minimal ventilatory support requirements and for those who are brain dead.
Limited data have compared the two approaches:
●A randomized study of 168 patients compared one-step ventilator cessation with a nurse-led algorithm in which ventilator support was weaned every 15 minutes for one hour before extubation during which patient comfort was measured using the respiratory distress observation scale and treated, if needed [13]. The weaning algorithm group had less respiratory distress and received fewer opioids and benzodiazepines.
●One prospective study of 458 patients compared gradual weaning with immediate extubation. In patients, immediate extubation was associated with more airway obstruction, gasping, and pain; in nurses, it was associated with less job stress [18]. Family/caregiver distress was similar between the groups. However, this study was not randomized; the increased rates of gasping in the immediate extubation group may have been due to selection bias or to how sedation and pain were managed before and during extubation. Patients in the immediate extubation group were more likely to be comatose whereas those in the gradual weaning group were more likely to have respiratory failure.
Gradual weaning — Using this approach, it is usually possible to decrease ventilator support over 10 to 60 minutes in a way that allows the patient to be comfortable breathing spontaneously on room air with minimal ventilator settings.
Our process is the following:
●We decrease or turn off telemetry and ventilator alarm settings and pulse oximetry.
●We decrease mechanical ventilatory support settings in a stepwise fashion (eg, every 10 to 15 minutes). With each ventilator change, we assess and treat signs of respiratory distress or pain before proceeding to the next step of ventilator weaning.
•We first decrease the positive end-expiratory pressure to zero and assess for respiratory distress.
•We next reduce the fraction of inspired oxygen (FiO2) by 20 percent decrements to 0.21.
•We next change the ventilator mode to pressure support (PS; eg, 8 to 10 cm H2O) and reduce in 1 to 2 cm decrements to PS 5 cm H2O.
●For patients with symptoms that arise during weaning, we administer bolus doses of intravenous analgesics (eg, fentanyl, morphine) as appropriate before any new change in ventilator settings (table 1).
There are no precise guidelines to facilitate dosing. In general, most clinicians base dosing on the patient's current doses of sedative or analgesics and the degree of distress.
The required opiate doses for pain and dyspnea vary widely from patient to patient (eg, morphine 1 to 20 mg intravenous push or even higher for patients with significant opioid tolerance). For example, patients who have been on opioids require higher doses to control symptoms relative to patients who have not been receiving opioids. There is no upper limit on acceptable doses as long as the dose is titrated to the patient's symptoms. A similar principle exists for intravenous benzodiazepines when used to alleviate anxiety (eg, midazolam 1 to 2 mg intravenous push).
Patients are typically already receiving continuous sedative infusions, so bolus doses are administered in addition to the infusion. Generally, we do not increase infusion rates during gradual weaning but rather rely on bolus dosing for a desired effect and adjust continuous infusion rates after extubation. (See 'Measures postextubation' below.)
●Once comfortable on low level ventilatory support (eg, PS 5 cm H2O), we extubate the patient. Exceptions to this include patients with severe upper airway edema, masses, or facial trauma that make it likely the patient would experience acute airway obstruction if the endotracheal tube (ETT) were removed. Excluding these circumstances, we favor extubation rather than leaving the ETT in place because doing so may allow for more patient comfort and facilitate a more natural dying process. The extubation procedure is similar to that for patients who undergo nonterminal extubation, except that the patient does not need to be instructed to breathe after ETT removal and oxygen does not need to be applied. This technique is described elsewhere. (See "Extubation management in the adult intensive care unit", section on 'Extubation equipment and technique'.)
Occasionally, patients experience acute respiratory distress after ETT removal due to unexpected upper airway obstruction. We generally manage this with rapid titration of short-acting intravenous opioids. In rare circumstances, it may be necessary to temporarily apply a jaw thrust maneuver (by placing the fingers behind the angle of the jaw and bringing the jaw forward) until the analgesics adequately take effect.
●Immediately after extubation, we assess for and treat symptoms of respiratory distress, if present. We remove the ventilator from the room. If family/caregivers/surrogates were not present for extubation, we invite them to come back inside the room. We then proceed to postextubation care. (See 'Measures postextubation' below.)
One-step weaning (immediate extubation) — One-step weaning involves the following:
●We decrease or turn off alarm settings and pulse oximetry.
●If the patient is on continuous infusions of opioids or sedatives, we continue those at the rate required to achieve patient comfort on the current ventilator settings before extubation. If the patient is on intermittent dosing of analgesics and sedatives, we ensure that the patient is comfortable before extubation.
●We consider whether additional anticipatory doses of intravenous opioids are needed before extubation. The rationale for this is that some patients may foreseeably develop significant respiratory distress following immediate extubation. For example, patients with intact consciousness who require high levels of ventilatory support or high FiO2 will often experience significant respiratory distress with immediate extubation. Anticipatory dosing of opioids is ethically appropriate to prevent or mitigate distress that is likely to occur immediately following extubation rather than subjecting the patient to experience discomfort before administering opioids.
●We extubate the patient following the last steps described above. (See 'Gradual weaning' above.)
Predicting time to asystole — Predicting the time to asystole helps set family/caregiver/surrogate expectations, plan resource utilization, and identify patients suitable for organ donation after circulatory death. For some patients, death will occur within a short amount of time (minutes to hours); for others, death may not occur for several days or more [19]. Rapid inability to tolerate a spontaneous breathing trial is a helpful indicator that a patient may die relatively quickly after withdrawal of mechanical ventilation. However, accurate prediction is challenging.
In a meta-analysis of 23 studies, variables associated with imminent asystole (ie, <60 minutes) included the following [20]:
•Deteriorating oxygenation
•Absent corneal reflexes
•Absent cough reflex
•Vasopressor use
•Comfort medication use
While predictive models have been proposed [21,22], further research and validation are needed before they can be routinely used.
MEASURES POSTEXTUBATION —
After extubation, we document the date and time of the procedure (ie, withdrawal of mechanical ventilation), how the procedure was tolerated, and what medications were needed. Postextubation care is outlined in the sections below.
Discontinue interventions that do not provide comfort — We typically discontinue the following (see "Palliative care: The last hours and days of life", section on 'Management of the dying patient'):
●Routine evaluations that provide no comfort to the patient (eg, blood draws and radiographs).
●Supplemental oxygen.
●Routine monitoring of blood pressure, temperature, and oxygen saturation. These vital signs add little to the clinician's bedside assessment of distress in dying patients and may cause family members to pay more attention to the monitors and alarms than the patient. However, we acknowledge that some families/caregivers/surrogates may prefer to have some monitoring continue, and this should be accommodated. To aid in symptom control, we often keep telemetry leads on the patient to allow monitoring for tachycardia and tachypnea. Many telemetry systems have the ability to turn off the monitor in the patient's room while still allowing visibility on monitors outside the room.
●Medications, treatments, and devices that do not promote patient comfort (eg, antibiotics, vasopressors, pacemakers [if consented by the family/caregiver/surrogate], artificial nutrition and hydration, intra-aortic balloon pumps, extracorporeal membrane oxygenation devices, dialysis). Some clinicians and family members prefer to leave urinary or rectal catheters in place to avoid the patient soiling themselves and the discomfort of cleaning. We also typically continue anticonvulsants as a comfort measure.
Continue to observe, assess, and treat symptoms in the ICU — Patients generally remain in the intensive care unit (ICU) until their symptoms are well-controlled and are not rapidly changing. Most patients undergoing withdrawal of ventilation in the ICU die in the ICU [23]. However, transition to a private room sometimes occurs when time to death is expected to be prolonged (eg, >24 hours). (See "Palliative care: The last hours and days of life", section on 'Place of death'.)
We continue frequent bedside assessments (every 15 minutes) and administer intravenous opioid boluses until the patient is reliably not showing signs of distress. Once stable, assessments can be lengthened.
Thereafter, we transition from bolus dosing to continuous infusions of opioids. If infusions are newly initiated, we typically start at half of the bolus dose required to alleviate symptoms (eg, if morphine 12 mg alleviated symptoms, the infusion can be initiated at 6 mg/hour and titrated up or down as needed). Anxiety should be treated with intravenous benzodiazepines. Rarely, low-dose phenobarbital or propofol are used.
The order for titrating medicines should not indicate titration to a vital sign, and certainly not one incompatible with life (eg, do not order, "titrate to achieve respiratory rate < breaths per minute"). Instead, we write orders that base titration on patient symptoms (eg, "For pain or respiratory distress, bolus 50 percent of the hourly dose and increase infusion rate by 1 mg/hour").
Symptom management during end-of-life care outside the ICU is described separately. (See "Palliative care: The last hours and days of life", section on 'Management of specific symptoms'.)
For patients in whom uncontrollable symptoms are present after extubation, palliative sedation may be needed to provide an appropriate level of comfort. (See "Palliative sedation", section on 'Indications'.)
Other supportive measures — Other supportive measures include the following:
●Frequent updates – We encourage families/caregivers/surrogates to spend time with their loved one after ventilator withdrawal, and we provide support strategies to help families cope with the dying process (eg, quiet space, soft lighting). We provide frequent updates to family/caregivers/surrogates with each assessment until death. Often refreshments are also provided during this process.
●Bereavement support – We offer bereavement support in the form of spiritual counseling, soft music, observing silence, or information packets. Personnel skilled in social work and case management are available to assist families and loved ones with the logistics of death.
Palliative care teams also help to provide bereavement follow-up or link with other services that provide this. They can also assist with symptom management and, in some settings, directly provide inpatient care. (See "Benefits, services, and models of subspecialty palliative care".)
Occasionally, staff members also require support.
●Condolences – The clinician and other allied health professionals express condolences after death.
SPECIAL POPULATIONS
Patients on prolonged mechanical ventilation — For patients requiring ventilatory support outside of the intensive care unit (eg, patients in long-term acute care hospitals with tracheostomies), ventilator withdrawal is similar to that described above (see 'Practical aspects of withdrawing mechanical ventilation' above). Removing the inner cannula of a tracheostomy tube is unnecessary, but the tracheostomy can be disconnected from the ventilator if preferred or if secretions are an issue.
The topic of prolonged mechanical ventilation is discussed separately. (See "Management and prognosis of patients requiring prolonged mechanical ventilation in long-term acute care hospitals (LTACH)".)
Massive hemoptysis — For patients with significant hemoptysis, maintaining an artificial airway rather than removing the endotracheal tube may be useful to provide access for palliative suctioning.
Patients receiving noninvasive ventilation — For patients receiving noninvasive ventilation (NIV) who require high levels of ventilatory support, we often perform a rapid weaning of support paired with treatment of respiratory distress, similar to that described for patients undergoing gradual endotracheal weaning (see 'Gradual weaning' above). For patients with minimal NIV settings (eg, 5 cm H20), we often immediately remove the mask without weaning the pressure, after having given an anticipatory dose of opioids as described for patients undergoing immediate extubation. (See 'One-step weaning (immediate extubation)' above.)
SOCIETY GUIDELINE LINKS —
Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Palliative care" and "Society guideline links: Intensive care unit ethics".)
SUMMARY AND RECOMMENDATIONS
●Goals of care – Discussions about patients' goals of care and preferences around invasive mechanical ventilation should ideally begin before patients develop impending respiratory failure.
•Details regarding the content of such discussions are provided separately:
-(See 'Goals of care discussions' above.)
-(See "Discussing goals of care".)
•It is essential that patients and families/caregivers/surrogates understand that ventilator withdrawal is not equivalent to withdrawal of patient care, that it is beneficial by being consistent with the patient's wishes, and that clinicians have a moral and ethical obligation to treat pain and other symptoms in a dying patient, even if doing so might hasten their death. (See 'Ethical issues specific to withdrawing ventilatory care' above.)
●Practical aspects – Our approach is outlined in the table (table 1).
•Each institution should develop protocols to guide a multidisciplinary approach to ventilator withdrawal, which can be further tailored to the patient's needs. (See 'Protocolized and multidisciplinary approach' above.)
•We explain the steps involved to family/caregivers/surrogates who are encouraged to gather together, spend time with the patient, meet with support staff, and indicate to the team when they are ready. (See 'Preparing the family or caregivers' above.)
•We ensure that neuromuscular blocking agents are discontinued well in advance and, if required, administer an agent for secretion control. We ensure a do-not-resuscitate/intubate (DNR/I) order is in place, intravenous access is secured, suctioning equipment is available, rapid-acting intravenous analgesics or sedatives are readily accessible, and potential continuous infusions are anticipated. (See 'Preparing the patient' above.)
•There are two approaches to ventilator withdrawal (see 'Choosing among the approaches' above):
-For most patients, we suggest a gradual approach followed by terminal extubation rather than one-step extubation (Grade 2C). This involves a stepwise reduction in ventilatory settings with assessment and treatment of symptoms before each ventilator change. This approach has been shown to cause less patient distress and may be more valuable for patients requiring high ventilator support. (See 'Gradual weaning' above.)
-Immediate extubation is an alternative and may be suitable for patients with severe neurologic injury and minimal sedative needs. When chosen, we pay particular attention to anticipatory dosing of sedatives and analgesics. (See 'One-step weaning (immediate extubation)' above.)
•For most patients, death will occur within a short amount of time (minutes to hours), although accurate prediction is challenging. (See 'Predicting time to asystole' above.)
●Postextubation – We discontinue the agreed-upon unnecessary therapies and monitoring (eg, blood draws and oxygen, blood pressure, temperature, oxygen saturation measurement) and medications or devices that do not promote comfort (eg, antibiotics, vasopressors, artificial nutrition and hydration, intra-aortic balloon pumps, extracorporeal membrane oxygenation devices, dialysis). We continue to treat symptoms and provide updates and emotional support to the family/caregivers/surrogates. Symptom-based palliation often necessitates a continuous infusion of opioids. (See 'Measures postextubation' above.)
ACKNOWLEDGMENT —
The UpToDate editorial staff acknowledges John Kellum, MD, who contributed to earlier versions of this topic review.
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟
