1. Loss of upper airway protective reflexes

2. Decreased level of consciousness

3. Effects on the central respiratory drive by hypoventilation syndromes or metabolic acid-base disturbances

Loss of Upper Airway Protection

After extubation, some patients may be at risk of clinically significant aspiration or failure to clear their respiratory secretions. Both can prevent patients from being able to safely maintain spontaneous, independent breathing. Before extubation, patients should be assessed for the presence of adequate cough and gag reflexes and their ability to cough well enough to clear their secretions (Table 4.3). Although unsettled, there are data that some patients without a gag reflex can still be successfully extubated. If patients lack a good cough, however, and are unable to clear secretions by themselves, weaning can continue but extubation should be delayed. If, with repeated testing over several days, the patient cannot adequately protect the airway or clear tracheal secretions, the patient should get an elective tracheostomy. This allows a secure access to the airways for suctioning secretions and is more comfortable than continuation of an endotracheal (ET) tube. A tracheostomy is also useful for monitoring aspiration of oral fluids during assessment of swallowing function and usually enhances the patient’s ability to communicate (Chapter 22).

The presence of swallowing dysfunction should be assessed in all recently extubated patients before oral intake. This is especially important in patients with poor or absent gag reflexes, patients with tracheostomies, or those with a history of suspected aspiration events (Chapter 22).

Decreased Mental Status

Intensive care unit (ICU) patients commonly have a decreased level of consciousness, often because they are sedated with medications. Level of consciousness influences attempts at weaning and extubation in two ways. First, a depressed mental status may result in loss of upper airway protection, as discussed earlier. Second, patients with decreased mental status may also have a decreased respiratory drive. As long as respiratory drive is considerably impaired, assisted ventilation is necessary. However, when the lack of respiratory drive is due to the absence of chemical stimuli to breathe (i.e., a low PaCO₂ with high pH), a tapering of assistance in breathing is reasonable and often effective. For example, well-oxygenated patients who are alkalemic because of iatrogenic hyperventilation while on mechanical ventilation may not breathe until their PaCO₂ levels are allowed to return to normal.

Several pivotal trials have demonstrated that strategies that minimize sedation can positively influence key patient outcomes, including ventilator weaning. Daily interruption of sedation—with sedatives held in mechanically ventilated patients until they are wakeful—has been found to reduce the average number of days patients spend requiring assisted ventilation. Additional approaches to minimize sedation, including the use of shorter-acting agents or administration of sedative medications in bolus form, especially if done according to a goal directed sedation protocol (carried out by the ICU nursing staff) as opposed to continuous intravenous (IV) infusions, have also been shown beneficial in weaning (Chapter 5). Furthermore, combining the practice of daily sedation interruptions with spontaneous breathing trials (SBTs) resulted in a greater number of ventilator-free days in mechanically ventilated medical patients in a multicenter trial in the United States. Interruption of sedation, however, may increase the risk of self-extubation, and patients should be closely monitored as they resume consciousness.

In addition to medication-induced alterations in consciousness, it is important to assess for hypoventilation resulting from central sleep apnea (CSA), as this may also compromise weaning. CSA occurs as a result of impaired sensitivity to PaCO₂ or Pao₂ levels and can manifest in several ways, including as obesity hypoventilation syndrome (see Chapter 80) or periodic breathing (Cheyne-Stokes breathing). This is particularly problematic for patients ventilated with weaning modes that require them to initiate breaths. Patients with CSA may not respond to increasing hypercapnia with the expected elevation in respiratory rate or tidal volume. Moreover, these patients may appear comfortable during a weaning trial, despite an increased PaCO₂ level, which is only detected by measuring arterial or central venous blood gases. Not considering hypoventilation syndromes in the differential for difficult weaning may make patients seem unweanable when, in fact, they only require scheduled ventilatory support at night and with naps (Chapter 25).

In contrast to patients with hypoventilatory syndromes who do not breathe enough, some patients breathe too much. This includes some patients with brain stem strokes who present with marked tachypnea as a result of central neurogenic hyperventilation. This is challenging because, in general, adults cannot sustain persistent respiratory rates of more than 36 to 40 breaths per minute for too long before they develop respiratory muscle fatigue. Suppression of respiratory drive by high-dose opioids may be successful in such patients, allowing them to be weaned.

Metabolic Acid-Base Disorders

In addition to alterations in ventilation secondary to sedation or CSA, the CNS control of respiratory drive is also affected by both metabolic acidoses and alkaloses. Normally, patients compensate for a metabolic acidosis (serum HCO₃⁻ <18 to 20 mEq/L) by “blowing down” (compensating) their PaCO₂ to maintain their arterial pH in or near the normal range. This respiratory compensation is achieved by hyperventilating, which increases alveolar and minute ventilation (see Chapter 83), but ultimately it may exceed a person’s ventilatory capacity and lead to respiratory muscle fatigue. Thus, in addition to correcting the underlying cause of the metabolic acidosis, patients may require bicarbonate supplementation in order for them to wean.

Alternatively, patients with a severe metabolic alkalosis (serum HCO₃^– >45 to 55 mEq/L) may have an elevated arterial pH, which results in a decreased stimulus to breathe, even when their PaCO₂ levels increase (Chapter 83). This can promote development of hypercapnia, as respiratory compensation for the metabolic alkalosis. Hypercapnia, in these situations, may be mistakenly attributed to respiratory muscle fatigue, an assumption that can delay weaning.

A related problem occurs in those patients with COPD who have chronic CO₂ retention and develop a compensatory elevated serum bicarbonate level. If their PaCO₂ is normalized because of mechanical ventilation, their kidneys no longer need to compensate and their serum bicarbonate level may fall, albeit to a seemingly normal value. Although their arterial blood gas (ABG) may appear “normal,” prior to discontinuation of mechanical ventilation, these patients often fail extubation because of increasing respiratory acidosis, which causes dyspnea, tachypnea, and respiratory muscle fatigue. The preferred strategy for patients with COPD who have chronic CO₂ retention is to use ventilator settings that maintain their PaCO₂ levels at their baseline elevated values (see Appendix B). This approach also tends to result in sustained elevation of their serum HCO₃^– so that, when the patient’s pulmonary function has returned toward its baseline, weaning has a reasonable chance to be successful.

Respiratory Muscle Weakness

In some patients, respiratory muscle weakness occurs as a primary event (e.g., in some neuromuscular disorders; see Chapter 67). In other patients, decreased respiratory muscle strength may occur secondary to the effects of critical illness or respiratory failure (see Chapter 48 and Table 4.2), such as when fatigued respiratory muscles need additional time to recover (which may take up to 1 day). Moreover, with muscle disuse—as occurs with modes of mechanical support that do not require significant engagement of the patient’s respiratory muscles—or protein malnutrition, the muscles atrophy. Indeed, evidence of histologic myofiber atrophy has been demonstrated in the diaphragm of humans mechanically ventilated for less than 1 day, possibly because of altered proteolysis. In response to concerns regarding ICU-acquired weakness, several clinical trials were found that, compared to usual care, daily physical and occupational therapy coupled with sedative interruptions resulted in earlier weaning, as evidenced by a greater number of ventilator-free days and a larger number of patients who regained their baseline functional independence (Chapters 5 and 21) .

Decreased muscle function may also be exacerbated by metabolic disorders such as hypophosphatemia and hypokalemia. Furthermore, severe hyperinflation—from airflow obstruction and auto–positive end-expiratory pressure (auto-PEEP)—compromises the efficiency of the length-tension relationship of the normally situated diaphragm and exacerbates the effects of muscle weakness.

Changes in Chest Bellows Function

Various factors can limit lung or chest expansion. In postoperative states, both the architecture of the incisions as well as the degree of pain may limit expansion. Prior to modern pain management strategies (see Chapter 87), on the first postoperative day, a patient’s vital capacity often decreased to only ~25% of preoperative values after thoracotomy as well as after upper abdominal surgeries (e.g., open cholecystectomy). Other common factors that can alter expansion include the presence of a flail chest (resulting from trauma or closed chest cardiac resuscitation), pleural effusions, major atelectasis as well as abdominal distention (from air, ascites, dialysate, or edema). Similarly, intra-abdominal hypertension and abdominal compartment syndrome (Chapters 90 and 97)—often a result of extensive fluid resuscitation—can limit diaphragmatic excursion, thereby decreasing forced vital capacity (FVC) and increasing dead space ventilation.

Upper Airway Injuries

Both iatrogenic and noniatrogenic injuries can result in upper airway obstruction leading to respiratory failure requiring initiation of mechanical ventilation. Patients who experienced difficult intubations—either for anatomic or situational reasons—may develop temporary vocal cord edema or permanent injury. Likewise, patients who undergo head and neck surgery may suffer from direct vocal cord injury. Smoke inhalation can result in both thermal and chemical injuries to the upper airways, sometimes with delayed manifestations of edema and sloughing that can limit air passage. Additionally, patients who develop angioedema—either secondary to an autoimmune disease or after exposure to medications—need adequate time for swelling to improve before successful extubation can occur.

Respiratory failure can occur from acute upper airway injury because of a resulting decreased ventilatory capacity—from upper airflow obstruction—and a compromised force-tension relationship from hyperinflation (flatting the domes of the diaphragms). In addition, as a rule, ventilatory load is increased by an increased resistive work of breathing and tachypnea.

Depending on the nature of the upper airway injury, it may take several days or more before patients are candidates for successful extubation. During this time, if the patient can tolerate it, he or she should be mechanically ventilated with as little support as is necessary to keep the patient breathing comfortably. Modes of ventilation that require patients to utilize their own respiratory muscles (e.g., pressure support [PS]) are often favored so as to limit muscle disuse atrophy that may occur when patients are ventilated with assist control (AC) modes that provide full support.

Two approaches have been used to determine if upper airway edema has sufficiently improved for successful extubation to occur. The first method involves direct visualization of the upper airway and vocal cords with a bronchoscope or a laryngoscope while the ET tube is in place. However, visualization of the laryngopharynx and vocal cords is often limited by the presence of the ET tube plus a feeding tube. In some cases, patients can also be extubated over a bronchoscope, which may allow a brief additional assessment of vocal cord movement as the ET tube is removed. Extubating a patient for whom one has high concern for upper airway patency should be done in coordination with an anesthesiologist, an otorhinolaryngologist, or both at the bedside.

An alternative approach to assess upper airway patency is to perform an “air leak” test—that is, to measure the air leak produced by deflation of the ET tube cuff—as the patient receives a defined tidal volume (e.g., 500 mL) during volume cycled mechanical ventilation. In patients with high risk for postextubation stridor, if the patient’s air leak around the deflated cuff is nil or modest (<100 mL per breath of 500 mL), one can assume that supraglottic obstruction may be present and delay extubation until after treatment of presumed laryngeal edema (i.e., IV dexamethasone and sitting the patient more upright in bed). However, because prospective studies have reported false positives (i.e., patients who had decreased air leaks but were extubated successfully) when applied to general ICU patients, some clinicians elect to limit its use to selected high-risk patients.