Show Summary Details
Page of

Discontinuing Mechanical Ventilation 

Discontinuing Mechanical Ventilation
Discontinuing Mechanical Ventilation

John W. Kreit

Page of

PRINTED FROM OXFORD MEDICINE ONLINE ( © Oxford University Press, 2020. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

date: 12 July 2020

So far, you’ve read a great deal about when to start and how to manage mechanical ventilation. Now we’re going to discuss when and how to remove or “wean” patients from the ventilator. While this may not seem as challenging (or as exciting) as endotracheal intubation and selecting ventilator settings, the process of discontinuing mechanical ventilation is just as important. That’s because unnecessary delays increase the incidence of hospital-acquired infections, lengthen ICU stay, and increase health care costs. On the other hand, stopping mechanical ventilation too soon exposes the patient to the risks associated with recurrent respiratory failure and re-intubation.

This chapter provides you with step-by-step instructions on how to discontinue mechanical ventilation. Although my goal is to help you avoid both delayed and premature extubation, as you’ll see, this process is often more of an art than a science, and no one gets it right every time. I’ll also provide an approach to the difficult-to-wean patient that emphasizes the identification and treatment of the underlying cause(s) of ventilator-dependence.

The Weaning Process

As shown in Figure 15.1, the process of discontinuing mechanical ventilation consists of four steps.

Figure 15.1 The four steps of the weaning process.

Figure 15.1 The four steps of the weaning process.

Step 1: Sedation Interruption

Sedation and analgesia are an essential and beneficial part of ICU care, but their benefits must be balanced by the fact that an impaired level of alertness delays weaning from mechanical ventilation. Studies have shown that routine, daily sedation-interruption allows the early identification of patients who are ready to resume spontaneous breathing and significantly decreases both the duration of mechanical ventilation and ICU length of stay.

Step 2: Determine Whether the Patient Is Ready to Begin the Weaning Process

This step requires that you understand why your patient required intubation and mechanical ventilation in the first place. As discussed in Chapter 7 and listed in Table 15.1, the reasons can be grouped into four major categories.

  • Acute or acute-on-chronic hypercapnia

  • Refractory hypoxemia

  • Inability to protect the lower airway

  • Upper airway obstruction

Table 15.1 Indications for Intubation and Mechanical Ventilation




Acute or acute-on-chronic hypercapnia

Reduced respiratory drive

Toxic/metabolic encephalopathy

Impaired transmission to the respiratory muscles

ALS, Guillain-Barre syndrome, myasthenia gravis, phrenic nerve injury

Respiratory muscle weakness

Metabolic myopathy, polymyositis

Severe lung or chest wall disease

COPD, asthma, pulmonary edema, morbid obesity

Refractory hypoxemia

Alveolar filling diseases

ARDS, pulmonary edema, pneumonia

Inability to protect the lower airway

Depressed level of consciousness plus vomiting risk

Toxic/metabolic encephalopathy plus upper GI bleeding or small-bowel obstruction

Upper airway obstruction

Anaphylaxis, epiglottitis, laryngeal mass, vocal cord paralysis

ALS = amyotrophic lateral sclerosis; COPD = chronic obstructive pulmonary disease; ARDS = acute respiratory distress syndrome; GI = gastrointestinal

Every day, you need to assess how much improvement there has been in the disorder that led to intubation. If your therapy has been largely unsuccessful, there’s no reason to think that your patient can be extubated. If, on the other hand, the original problem has resolved, or has at least significantly improved, your patient is probably ready to start the weaning process. While sedation is on hold, check to see whether your patient meets the general criteria shown in Box 15.1. If so, it’s time to move on to the next step.

FIO2 = fractional inspired oxygen concentration; PEEP = positive end-expiratory pressure

Step 3: Perform a Spontaneous Breathing Trial

Now it’s time to see whether your patient can breathe on their own by performing a spontaneous breathing trial (SBT). This can be done using either a “T-piece” or low-level pressure support (PS) breaths, and these two methods are compared in Table 15.2. During a T-piece trial, the patient is removed from the ventilator, and the endotracheal tube is connected at a right angle to large-bore tubing (hence the name “T-piece”) that carries high-flow, humidified gas. During a pressure support trial, the patient remains connected to the ventilator but only receives a small amount of inspiratory assistance (usually 5 cmH2O), which helps to overcome the extra viscous forces created by the endotracheal tube and ventilator circuit.

Table 15.2 Comparison of T-piece and Pressure Support Spontaneous Breathing trials


Pressure Support

Connected to ventilator



Inspiratory pleural pressure during breathing trial



Provides supplemental oxygen



Compensates for extra viscous forces



Can add PEEP



Able to measure tidal volumes



Ventilator alarms active



PEEP = positive end-expiratory pressure

As you can see from Table 15.2, there are lots of advantages to using PS trials. Most importantly, since the patient remains connected to the ventilator, you can easily monitor respiratory rate and tidal volume, and an alarm will sound when preset limits are reached. This gives PS trials a huge safety advantage. Pressure support trials also allow the patient to be suctioned more easily and eliminate the need (and the time required) for a respiratory therapist to repeatedly disconnect and reconnect the ventilator.

In fact, I use T-piece trials only when patients have severe left ventricular (LV) dysfunction. As discussed in Chapter 3, the increase in pleural pressure produced by a switch from spontaneous to mechanical ventilation actually assists the failing heart by reducing both LV preload and afterload. Patients with systolic heart failure may do well on a PS trial but then develop acute pulmonary edema after extubation. A T-piece trial subjects the heart to the same pleural pressures and the same preload and afterload that will be present after extubation. In this way, occult myocardial ischemia or decompensated LV failure usually declares itself prior to removal of the endotracheal tube.

Step 4: Determine If the Patient Is Ready for Extubation

Although there’s no foolproof method of accurately predicting whether extubation will be successful, making sure that your patient meets all of the following criteria during a SBT will allow you to be right almost all of the time:

  • The patient is awake, alert, and cooperative.

  • The patient’s spontaneous respiratory rate is ≤25 breaths per minute.

  • The patient’s spontaneous tidal volume is ≥300 ml.

  • When asked, the patient is able to significantly increase his/her spontaneous tidal volume.

Blood gas measurements are not routinely needed prior to extubation. I do, however, measure pH and PaCO2 when patients have chronic hypercapnia. I’ve learned the hard way that patients with a rising PaCO2 can look very comfortable during an SBT.

There are two important questions that have not been addressed. The first is: How long should an SBT last? With great confidence, I can answer: “It depends.” For example, a patient who is alert after being intubated for a drug overdose usually needs an SBT of no more than 5–10 minutes. The same is true for a healthy postoperative patient who has fully awakened from general anesthesia. On the other hand, most critical care physicians prefer a longer observation period (usually ½–1 hour) when patients have chronic lung or cardiac disease, neuromuscular disease, morbid obesity, or other disorders that predispose them to recurrent respiratory failure. Continuing a trial beyond one hour does not provide additional prognostic information. The second question is much easier to answer: What do you do when a patient does poorly on an SBT? Simply return them to their previous ventilator settings and try again in 12–24 hours.

Other Considerations

Once a patient has “passed” an SBT, you need to consider a few more issues before proceeding with extubation:

  • Is the patient at risk for post-extubation laryngeal edema?

  • Will the patient be able to effectively clear secretions from the airways following extubation?

  • Does the patient have a “difficult airway” should the need for re-intubation arise?

Laryngeal Edema

The likelihood of post-extubation laryngeal edema and upper-airway obstruction increases with the duration of intubation. Prior airway trauma and multiple intubations are other important risk factors. Because laryngeal edema is very difficult to detect while the endotracheal tube is in place, a “cuff leak test” is often performed in patients who are at risk for this complication. This is done by deflating the endotracheal tube cuff and listening for the sound of an air leak during a mechanical breath. The theory is that significant laryngeal edema will prevent air from escaping even when the trachea is no longer sealed. Although this makes a lot of sense, the absence of an air leak is, unfortunately, fairly nonspecific. That is, many (most) patients without an air leak will have no sign of upper airway obstruction after extubation. Nevertheless, in high-risk patients, the absence of an air leak tells you that you must be ready to deal with upper airway obstruction following extubation.

Effective Airway Clearance

Patients must have an effective cough if they are to clear secretions after the endotracheal tube is removed. Although this is difficult to assess while they are intubated, the answers to these three questions will help you predict how a patient will do after extubation:

  • What is the patient’s level of alertness?

  • How strong is the patient’s cough during suctioning?

  • How often does the patient need to be suctioned?

For example, a patient who is wide awake, has a vigorous cough with suctioning, and has minimal airway secretions will almost certainly have no difficulty maintaining airway clearance following extubation. On the other hand, a patient who is very lethargic, requires frequent suctioning, and has little or no cough response, should not be extubated even if they do well on an SBT. Of course, there’s a large gray area between these two extremes, but the answers to these questions will at least give you a framework for judging the advisability of extubation.

The Difficult Airway

A patient is said to have a “difficult airway” when they cannot be adequately ventilated with a bag-mask, intubated using standard laryngoscopy, or both. This can result from a number of factors, including morbid obesity, a recessed mandible (retrognathia), and temporo-mandibular joint or cervical spine immobility. Critical care physicians are trained to screen for these problems prior to intubation, but it’s essential that this assessment also be performed before extubation. Information about prior intubation attempts is, of course, invaluable, so be sure to review the medical record and speak with your colleagues. This pre-extubation assessment is important because patients with a difficult airway have a high risk of morbidity and even death if re-intubation is required. So make sure that you have all the equipment and personnel needed before extubating a patient with a known or suspected difficult airway.

Sir William Osler once said, “Medicine is an art of probability,” and this certainly pertains to the process of determining whether or not a patient should be extubated. Despite your best clinical judgment and lots of experience, some patients who are judged ready for extubation will subsequently need to be re-intubated. There are undoubtedly patients in whom the opposite is true. For your reference (and consolation), most studies report extubation failure rates of 5–15%. It’s often said that if you are not re-intubating a few patients, you probably are not extubating soon enough!

Approach to the Difficult-to-Wean Patient

Unfortunately, many patients cannot be quickly removed from mechanical ventilation. This may be because the underlying reason for intubation has failed to improve or because the patient has developed one or more new problems while on the ventilator. Often, both occur. When faced with a patient who consistently does poorly during SBTs, your goal is to determine the reason(s) for ventilator-dependence and then come up with a therapeutic strategy.

To begin this process, it’s very useful to think of ventilator-dependence as being the result of an abnormal relationship between ventilatory ability and ventilatory demand (Figure 15.2). Ventilatory ability is the minute ventilation(V˙E) that a patient can maintain for an indefinite period of time. Ventilatory demand is the V˙E needed to excrete CO2 and maintain an acceptable PaCO2 and arterial pH. Normally, of course, ability far outstrips demand. That’s why breathing is usually so effortless. Ventilator-dependence occurs when a patient’s ventilatory demand exceeds their ability. This may be caused by increased demand, impaired ability, or (often) both.

Figure 15.2 (A) Normally, ventilatory ability far exceeds ventilatory demand.(B) Ventilator-dependence occurs when demand exceeds ability.

Figure 15.2 (A) Normally, ventilatory ability far exceeds ventilatory demand.

(B) Ventilator-dependence occurs when demand exceeds ability.

Like routine weaning, the evaluation and management of the ventilator-dependent patient can be broken down into several steps (Figure 15.3).

Figure 15.3 The three steps for evaluating and treating the difficult-to-wean patient.

Figure 15.3 The three steps for evaluating and treating the difficult-to-wean patient.

Step 1: Assess Ventilatory Ability and Demand

Ventilatory demand is determined by noting the V˙E needed during assisted mechanical ventilation to generate an “appropriate” PaCO2. For most patients, of course, this is around 40 mmHg, but it may be either higher (e.g., in a patient with chronic hypercapnia) or lower (e.g., during compensation for a metabolic acidosis). If you keep in mind that V˙E is normally in the range of 6–8 L/min, you can get a good sense of whether your patient’s requirements are excessive. Next, measure V˙E during totally unassisted breathing using a pressure support level of zero. Now compare these two numbers. In patients who have repeatedly failed attempts to discontinue mechanical ventilation, demand will almost always exceed ability, and it will be obvious why the patient cannot maintain spontaneous ventilation.

Step 2: Determine the Cause(s) of Increased Demand, Impaired Ability, or Both

Table 15.3 lists the causes of impaired ventilatory ability and excessive demand. Carefully evaluate your patient. Can you explain their abnormal ability–demand relationship? In most cases, patients have several reasons for ventilator-dependence. I’ll now discuss the most common causes.

Table 15.3 Causes of Ventilator-Dependence




Reduced Ability

Neuromuscular disease

Altered mental status

Over-sedation, delirium, encephalopathy

Inadequate respiratory drive

Over-sedation, delirium, encephalopathy

Impaired conduction to the respiratory muscles

Critical illness polyneuropathy

Respiratory muscle weakness

Critical illness myopathy

Lung disease

Obstructive lung disease

COPD, asthma, bronchiectasis

Restrictive lung disease

Interstitial lung disease, organizing lung injury, interstitial edema

Air space filling diseases

Pulmonary edema, pneumonia, atelectasis

Chest wall disease

Morbid obesity

Large pleural effusions

Massive ascites

Cardiovascular disease

LV failure

Cardiomyopathy, ischemia, diastolic dysfunction

Increased Demand

Metabolic acidosis

Renal failure, dilutional acidosis, diarrhea


Pain, anxiety, iatrogenic

Elevated dead space ventilation

Significant underlying lung disease

Hyper-metabolic states

Fever, infection, hyperthyroidism

Reduced Ability

Over-Sedation, Delirium, and Encephalopathy

These three conditions are extremely common in ICU patients and can have a profound effect on our ability to wean them from mechanical ventilation. Regardless of the cause, any process that impairs alertness is likely to reduce spontaneous tidal volume. Spontaneous respiratory rate may also fall, but it typically increases to compensate for the drop in tidal volume.

Critical Illness Polyneuropathy and Myopathy

Clinically significant muscle weakness due to critical illness polyneuropathy (CIP), critical illness myopathy (CIM), or both, is believed to occur in as many as one-third of mechanically ventilated patients. Since involvement of the diaphragm and the other respiratory muscles is common, these disorders are a very important cause of ventilator-dependence. Although their pathogenesis remains unclear, the most important risk factor for CIM is the use of corticosteroids and neuromuscular blocking agents, whereas CIP occurs primarily in patients with the systemic inflammatory response syndrome (SIRS) or severe sepsis.

Since isolated diaphragm weakness is rare, patients with persistent ventilator-dependence can be screened for CIP and CIM simply by assessing their peripheral muscle strength, and the diagnosis can usually be confirmed with electromyography and nerve-conduction studies. The treatment of CIM and CIP is supportive and both conditions usually improve over a period of weeks to months.

Lung Disease

Ventilator-associated pneumonia and atelectasis are common in patients with respiratory failure and may cause or contribute to ventilator-dependence. Unfortunately, these disorders are often unrecognized because they may not be evident on a portable chest radiograph. Computed tomography, with its significantly greater sensitivity and specificity, can be very helpful in identifying or excluding these disorders. Also, make sure that you have optimized bronchodilator and anti-inflammatory therapy in patients with underlying obstructive lung disease.

Chest Wall Disease

Morbid obesity, pleural effusions, and ascites all reduce chest wall compliance, so during spontaneous breathing, much more inspiratory effort is needed to generate an acceptable tidal volume. Not surprisingly, this often leads to rapid, shallow breathing and repeated SBT failure. Draining large effusions and ascites can aid the weaning process. In morbidly obese patients, it is essential to perform SBTs in an upright position, either in the bed, or better yet, in a chair. This shifts weight off the chest wall, improves compliance, and increases tidal volume.

LV Dysfunction

Systolic and/or diastolic heart failure leading to pulmonary edema is another common, treatable, and often unrecognized cause of ventilator-dependence. Echocardiographic assessment of LV function is mandatory in all difficult-to-wean patients.

Increased Demand

Metabolic Acidosis

The respiratory system compensates for metabolic acidosis by increasing minute ventilation, which decreases the PaCO2 and increases the arterial pH. Although it’s well-tolerated in patients with normal lung function, this increase in required V˙E may lead to ventilator-dependence in patients with limited ventilatory reserve. Renal failure, diarrhea, and large infusions of crystalloid (“dilutional acidosis”) are the most common causes of persistent metabolic acidosis, and correcting the acidosis can be very helpful in some patients with ventilator-dependence.


Like respiratory compensation for metabolic acidosis, hyperventilation during mechanical ventilation causes the PaCO2 to fall, increases the required V˙E, and may impair the weaning process. The mechanism is a little different, though. Chronic hyperventilation and the resulting respiratory alkalosis lead to decreased renal reabsorption of bicarbonate and a fall in the serum bicarbonate concentration. During an SBT, the patient must then maintain the preexisting low PaCO2 in order to keep the arterial pH normal, and this increases ventilatory demand.

During mechanical ventilation, persistent hyperventilation may be caused by excessive patient-triggering in the setting of uncontrolled anxiety or pain. In such cases, providing adequate sedation or analgesia will usually correct the problem. Much more commonly, hyperventilation results from an inappropriately high set respiratory rate (iatrogenic hyperventilation). You should always suspect this when your patient’s total and set respiratory rate are the same. When this happens, simply reduce the set rate until the patient begins to trigger additional, spontaneous breaths.

As discussed in Chapter 13, it’s very important to remember that hyperventilation can cause not only a truly low PaCO2 (i.e., <40 mmHg) but also a PaCO2 that is less than normal for the patient. For example, a PaCO2 of 50 mmHg would be abnormally low for a patient with severe COPD and a baseline PaCO2 of 80 mmHg. If this low PaCO2 persists, the serum bicarbonate concentration will fall, and during an SBT, the patient will have a ventilation requirement that they couldn’t maintain even before they got sick and ended up on the ventilator. It should be obvious that such a patient has no chance of weaning from mechanical ventilation until both the PaCO2 and the serum bicarbonate concentration are restored to their baseline levels. How do you identify this problem? The best way is to find blood gas measurements that were done prior to the patient’s current hospitalization. If these aren’t available, review previous serum bicarbonate concentrations and make sure that there hasn’t been a significant change.

Increased Dead Space Ventilation

Recall from Chapter 2 that any disorder that affects the airways, parenchyma, or vasculature of the lungs causes both abnormally low and abnormally high ventilation-perfusion (V˙/Q˙) ratios. This leads to arterial hypoxemia, an increase in wasted or “dead space” ventilation, and an increase in the V˙E needed to maintain a given PaCO2. So, you can see that lung disease can not only reduce the sustainable V˙E, as discussed previously, it can also increase ventilatory demand. That’s why it’s so important to diagnose and effectively treat any type of lung disease when a patient is having difficulty coming off of mechanical ventilation.

Hyper-Metabolic States

Fever, uncontrolled infection, and hyperthyroidism increase total CO2 production. This increases the V˙E needed to maintain an appropriate PaCO2. Ventilatory demand often falls significantly once these problems have been recognized and treated.

Step 3: Treat the Causes of Increased Demand and Impaired Ability

The evaluation described in Step 2 will often identify not just one, but several potential causes of increased ventilatory demand and impaired ability. Your job is to now do everything you can to correct these problems. If you are able to restore a more normal relationship between demand and ability, your patient has a good chance of resuming spontaneous ventilation, and the weaning process should be continued. If the problems you’ve identified cannot be treated, or if demand still exceeds ability following therapy, you can conclude that the patient is likely to remain ventilator-dependent and that additional weaning attempts are unlikely to be successful.

When to Perform a Tracheostomy

As outlined in Box 15.2, tracheostomy has many advantages over trans-laryngeal intubation with an endotracheal tube. On the other hand, tracheostomy is a surgical procedure and carries a small risk of complications, including bleeding, infection, pneumothorax, and tracheal stenosis. There is nearly universal agreement that patients who have difficulty weaning from mechanical ventilation should undergo tracheostomy. When it should be done continues to be the subject of debate and controversy. That’s because the results of randomized, controlled trials of early (2–8 days) versus late (13–16 days) tracheostomy have been inconsistent and contradictory.

I favor an individualized approach. Patients who are alert, able to interact and communicate, and get out of bed are most likely to benefit from the advantages shown in Box 15.2. They should have a tracheostomy performed as soon as you are convinced that they will not be able to sustain spontaneous ventilation for at least a few weeks. Patients who are less alert and more debilitated will derive little benefit from early tracheostomy, and the procedure can be delayed, especially if the cause of respiratory failure is improving. When respiratory failure has been caused by a chronic disease (e.g., Guillain-Barre syndrome or spinal cord injury), there is no reason to delay the procedure, and regardless of patient characteristics, tracheostomy should be performed as soon as possible.

Additional Reading

Artime CA, Hagberg CA. Tracheal extubation. Respir Care. 2014;59:991–1005.Find this resource:

McConville JF, Kress JP. Weaning patients from the ventilator. N Engl J Med. 2012;367:2233–2239.Find this resource:

Perren A, Brochard L. Managing the apparent and hidden difficulties of weaning from mechanical ventilation. Intensive Care Med. 2013;39:1885–1895.Find this resource: