Patients with, or likely to develop, respiratory failure are the individuals who require MV support. Although there is overlap, respiratory failure can be generally categorized into lung failure and pump failure. Lung failure is pure gas-exchange failure and is manifested by hypoxemia. It is commonly due to the acute respiratory distress syndrome or cardiogenic pulmonary edema. Pump failure is synonymous with ventilatory failure and is manifested by hypercapnia and hypoxemia. It is commonly due to central nervous system depression (e.g., overdose, anesthesia) or respiratory muscle fatigue or weakness.

For those who recover from the insult that necessitated MV, most (80% to 90%) [1,2,3,4] can have MV easily discontinued and be extubated. In this group, MV can be discontinued in 77% of patients within 72 hours of the initiation of MV [4]. This group is composed predominantly of postoperative patients, patients with overdoses, and patients whose conditions cause pure lung failure that reverses rapidly. In the minority of patients, probably 10% to 20% overall, MV is more difficult to discontinue. Data suggest that duration of MV does not necessarily have an impact on long-term survival. For example, 1-year survival for patients on MV for more than 21 days can
be as high as 93% [5]. Although it may take 3 months or longer to be able to discontinue MV in these patients in long-term facilities, the ultimate quality of life of the survivors ranges from being minimally to moderately impaired [6,7].

There are potentially four separate and reversible reasons for prolonged MV [8] (Table 60.1).

Although no studies have been performed to determine systematically the relative importance of these factors, and combinations of these factors may be responsible for prolonged MV, the literature suggests that pump failure due to inspiratory respiratory muscle fatigue/weakness [12] is primarily responsible for failure of discontinuation of MV in these patients [3,13,14]. Muscle fatigue is “a condition in which there is loss in the capacity for developing force and/or velocity of a muscle, resulting from muscle activity under load and which is reversible by rest” [15,16]. Muscle weakness is “a condition in which the capacity of a rested muscle to generate force is impaired” [15,16]. Although fatigue and weakness can be experimentally distinguished, this is not usually possible in the clinical setting. Therefore, the term muscle fatigue, when used clinically and by us in this chapter, may actually encompass fatigue or weakness, or both. Contributors to respiratory muscle fatigue may be (a) central nervous system depression, (b) mechanical defects (e.g., flail chest and kyphoscoliosis) that increase the work of breathing, (c) lung disease that increases the work of breathing, and (d) mediators of ongoing active diseases (e.g., sepsis, ventilator-induced diaphragmatic dysfunction) that adversely affect the respiratory muscles.

The cause of inspiratory respiratory muscle fatigue is likely to be multifactorial [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36]. The major factors that compromise muscle strength and endurance are listed in Table 60.2. A few items deserve additional explanation.

Because there are no objective, rigorously generated data to determine the appropriate time to initiate weaning, physicians must rely on their clinical judgment. Therefore, the authors recommend that clinicians consider a carefully monitored spontaneous breathing trial (SBT) of discontinuation when the following criteria, set forth in a national clinical practice guideline, have been met: (a) The underlying reason(s) for MV has been stabilized and the patient is improving, (b) the patient is hemodynamically stable on minimal-to-no pressors, (c) oxygenation is adequate (e.g., PaO₂/FIO₂ greater than 200, PEEP no more than 7.5 cm H₂O, FIO₂ less than 0.5), and (d) the patient is able to initiate spontaneous inspiratory efforts [45]. Because potentially harmful effects of suddenly having to take on the work of breathing occur early (albeit infrequently) during SBTs [46], patients should be closely monitored during the first 5 minutes. Basing weaning decisions on the rapid shallow breathing index (RSBI) in effect enforces this (see section Predictive Indices for Total Discontinuation of Mechanical Ventilation). SBTs, variably performed with a T-piece, with low-level pressure support ventilation, or with just a predetermined amount of continuous positive airway pressure (CPAP) in ventilators equipped with “flow-by” internal circuits, should be timed to coincide with the daily sedation holiday to maximize the opportunity for success and to allow assessment of patient comfort and behavioral effects on breathing.

If the patient deteriorates or becomes distressed during this brief period of observation, MV should be reinstituted. The authors caution against assuming that anxiety is causing the failure of a breathing trial. Although anxiety can mimic respiratory failure, in the authors’ experience anxiety is not usually the cause of failure but rather a consequence of it. In fact, the “art of weaning” centers on the judgment whether weaning-induced distress is a manifestation of agitated delirium, sedative and narcotic withdrawal, pain and tube discomfort, or respiratory failure. When in doubt, the provider should assume the latter. We know of no validated test capable of distinguishing between these entities. To help decide in these situations, we sometimes observe patients who are difficult to wean while keeping them heavily sedated. If under these circumstances, unassisted breathing can be sustained without hypercapnia, hypoxemia, tachypnea, and tachycardia, we conclude that respiratory failure is no longer present, that agitation may be related to pain, anxiety, or sedative/hypnotic withdrawal, and proceed with a trial of extubation if and when we believe that the patient is able to protect his or her airway against the possibility of aspiration.

There are no data to show that attempts at starting the discontinuation of MV in this context lead to adverse consequences. On the contrary, screening patients daily to identify those who can breathe spontaneously can reduce the duration of MV and the cost of intensive care [47]. Because the authors’ recommendations are guidelines and not rigorously tested criteria, it may also be appropriate to start the carefully monitored process in an individual patient who has not met all of the previously mentioned guidelines.

Studies have evaluated a wide variety of physiologic indices to predict a patient’s ability to breathe spontaneously without MV [48]. These studies yield conflicting data due in large part to differences in methods and experimental design, such as population studied, choice of physiologic index threshold value, measurement techniques, definitions of success and failure, and perhaps because of selection bias in choosing patients for weaning studies [48].

A collective task force of clinician investigators co-facilitated by the American College of Chest Physicians, the American Association for Respiratory Care, and the American College of Critical Care Medicine developed evidence-based guidelines for weaning and discontinuing ventilatory support [45]. In their report, they evaluated the evidence for predicting success in weaning from MV [48]. A summary of their findings is as follows:

Although clinical observation of the respiratory muscles during spontaneous breaths was initially thought to be reliable in predicting subsequent discontinuation failure, respiratory-inductive plethysmographic studies [53] have shown this to be not necessarily the case. Any time there is a substantial increase in load on the respiratory muscles, a change in the rate, depth, and pattern of breathing may be observed. Because these signs may also be manifestations of fatigue, it is useful to note them. If these signs never appear, successful discontinuation is likely. If they do appear, patients must be observed closely for further deterioration because discontinuation inevitably fails if these signs are owing to fatigue.

Once MV has been discontinued, consider whether the patient is likely to fail extubation. The most common causes of extubation failure are upper-airway obstruction and inability to protect the airway and clear secretions. Patients at the highest risk of postextubation upper-airway obstruction are those who have been on prolonged MV, are female, and who have had repeated or traumatic intubations [54]. One method of assessing for the presence of upper airway obstruction during MV is the cuff-leak test. It is performed by comparing the exhaled volumes before and after the balloon of the endotracheal tube has been deflated. Although one study [55] showed that a cuff leak of less than 110 mL measured during assist-control ventilation within 24 hours of extubation identified patients at high risk of postextubation stridor, other studies have not [56]. Although the concept of measuring cuff leak is intuitively appealing, the benefits are not clearly identified, and the process and even the actual values for decision making are not broadly agreed upon. Values of 110, 130, and 140 mL are all used in recent studies. Other studies use an approach of auscultation to detect leak. In addition, the appropriate course of action to take for an abnormal test is not defined. Some authors suggest treatment with steroids, some delay of extubation, and some advocate having persons with advanced airway skills present for the extubation. Therefore, because we are unable to scientifically determine which patients should have a test, how we would conduct the test, and what we would do with an abnormal value if we had one, we do not advocate routinely performing or basing decisions on the results of a cuff-leak test. A provider may consider using a cuff-leak test in specific patients to gain a general appreciation of the airway status in a high-risk patient [57,58,59].

Patients may also fail extubation because they are unable to protect their airways or clear their secretions. A prospective observational study [60] showed that the strongest predictors of extubation failure in patients who passed a SB trial were (a) poor cough defined as a cough peak flow measurement of less than 60 L per minute, (b) secretion volume of 2.5 mL per hour or greater, and (c) poor mentation as determined by the inability to complete any of the four following tasks on command: open eyes, follow observer with eyes, grasp hand, and stick out tongue. In this series, reintubation took place in 12% of patients when one of these predictors was present and 80% when all three were present. (See Chapter 62 for an in-depth discussion of cough effectiveness and how to assess for it.)

Once extubation has taken place, the authors proceed cautiously before instituting feedings by mouth. Because there is no clinically reliable way of assessing the adequacy of swallowing at the bedside, a formal swallowing evaluation (e.g., speech pathology consult and videofluoroscopic evaluation of swallow) should be considered in patients at increased risk of aspiration before resuming oral feedings. Although it is commonly appreciated that older age, debilitation, sedation, oral or nasal enteral feeding tubes, history of dysphagia, acute stroke, cervical spine surgery, muscle weakness, and/or tracheostomy are risk factors for aspiration, it is less commonly known that endotracheal intubation carries the same risk [61,62]. After extubation, swallowing difficulties may exist in up to 50% of patients for up to 1 week, even when endotracheal intubation has been of short duration, and the patient is awake and not seriously ill. In awake, postsurgical patients evaluated for aspiration following extubation, 50% of those who aspirated did so immediately when fed, whereas 25% and 5% aspirated when tested 4 and 8 hours later, respectively. (See Chapter 54 for an in-depth discussion of this subject.)

When the patient’s clinical condition has been stabilized, it is reasonable to consider starting the discontinuation process even if predictive index thresholds for success have not been met. Valuable time may be lost in liberating patients from the ventilator if one relies solely on these indices because they are not powerful predictors of success or failure. Furthermore, there is no evidence that shows that unsuccessful discontinuation trials have long-term adverse consequences, provided patients are monitored closely and certain pitfalls are avoided. For example, it is unwise to attempt SBTs on patients with active ischemic heart disease because systemic oxygen demand and cardiac output can increase substantially during transition from controlled MV to SB [63,64]. Patients must be prepared psychologically to understand that failing a discontinuation trial has no bearing on their ultimate prognosis. Finally, it is prudent to guarantee sufficient respiratory muscle rest after a failed attempt at SB. With few exceptions, such as patients recovering from general anesthesia or sedation with or without muscle paralysis, the authors usually do not have their patients undergo more than one (failed) discontinuation trial in any 24-hour period. This practice is supported by the work of Esteban et al. [65],
who showed that twice-daily SB trials offered no advantage over once-daily trials. Moreover, the inspiratory effort associated with a failed weaning trial may be sufficient to induce muscle fatigue that may not recover [65], unless it is followed by an extended period of rest.

With respect to extubation, it is reasonable to proceed when the patient’s ability to protect the airway suggests that extubation will be successful. We do not routinely administer systemic steroids to prevent postextubation stridor because of the inconsistent benefit seen in studies, and the uncertain timing of extubation encountered in clinical practice, which could potentially lead to extended courses of steroids with their associated side effects [66].