50 Weaning from Mechanical Ventilation
The Concept of Liberation and Extubation
Weaning from mechanical ventilation represents the period of transition from total ventilatory support to spontaneous breathing. About 70% of intubated mechanically ventilated patients are extubated on the first spontaneous breathing trial (SBT) attempt, whether by disconnection from the ventilator or after breathing at low levels of pressure support for short periods of time, such as 30 to 120 minutes.1,2 This pattern has recently been categorized as “simple weaning,” and the prognosis for such patients is good. The remaining patients, about 30%, need progressive withdrawal from artificial ventilatory support. These patients can be classified either as “difficult weaning” when they require up to three SBTs to achieve successful weaning, or “prolonged weaning” if they fail at least three weaning attempts or require more than 7 days of ventilatory support from the first SBT. The mortality rate for patients not simple/easy to wean is approximately 25%.3
Early liberation from mechanical ventilation and removal of the endotracheal tube is clinically important. Unnecessary prolongation of mechanical ventilation increases the risks of complications including infections (particularly of bronchopulmonary origin), barotrauma, cardiovascular compromise, tracheal injuries, and muscle deconditioning. To optimize patient outcomes, clinicians should hasten the process that ultimately leads to removal of the endotracheal tube.4
Liberation and extubation are different issues.5 Liberation refers to weaning from mechanical ventilation and means that a patient no longer requires ventilatory support. When this step is achieved, the clinician has to consider a different question: Is the patient able to breathe spontaneously without the endotracheal tube? Removal of the endotracheal tube is referred to as extubation. In terms of magnitude, the extubation failure rate—that is, the need to replace the endotracheal tube and reinstitute mechanical ventilation—is variable and ranges from 5% to 20% of extubated patients.1,6–8
Mechanisms Explaining Liberation Failure
Respiratory Pump Failure
The most common reason for weaning failure is respiratory pump insufficiency due to an imbalance between the patient’s capabilities and respiratory demands.9–11 During spontaneous breathing, the inspiratory muscles must generate sufficient force to overcome the elastance of the lungs and chest wall (lung and chest wall elastic loads) as well as the airway and tissue resistances (resistive load). This requires signal generation in the respiratory centers of the brainstem, anatomic and functional integrity of nerves that conduct the signal, unimpaired neuromuscular transmission, and adequate muscle strength (the aggregate term is neuromuscular competence). The ability of the respiratory muscles to sustain these loads without fatiguing is called endurance and is determined by the balance between energy supply and energy demand.
Jubran and Tobin12 investigated the progression of respiratory mechanics during SBT in patients with chronic obstructive pulmonary disease (COPD). At the very beginning of the trials, patients who subsequently failed had a slightly higher airway resistance, respiratory system elastance, and intrinsic positive end-expiratory pressure (PEEP) compared to those who succeeded. However, during the course of the trials, respiratory mechanics progressively worsened in patients who failed to be liberated from the ventilator. Subjects who failed developed rapid, shallow breathing, and most developed an increase in PaCO2. Together these abnormalities resulted in increased inspiratory muscle effort which, in some patients, was probably close to the threshold of muscle fatigue.
The issue of fatigue has been revisited by Laghi et al.13 The authors studied 19 intubated patients during weaning from mechanical ventilation. Eleven patients failed and eight succeeded. Several physiologic indices were measured before and 30 minutes after SBT. The transdiaphragmatic twitch pressure, elicited by magnetic bilateral phrenic stimulation, did not differ before the SBT between the patients that failed or succeeded at ventilator liberation, and this variable did not decrease after the trial in either group. A fall in transdiaphragmatic twitch pressure is a physiologic index of low-frequency fatigue. Patients failing the SBT were reconnected to the ventilator because of clinical signs of intolerance. These alterations, together with the reinstitution of mechanical ventilation, are mechanisms that might defend against the development of low-frequency fatigue. It was concluded that weaning failure was not accompanied by low-frequency diaphragmatic fatigue, although weaning-failure patients exhibited severe diaphragmatic weakness, since twitch pressures were always low.
Common Disorders That Alter the Balance of Capacity and Load in Critical Illness
Reduced Neuromuscular Capacity
Reduced output of the respiratory control centers may occur following administration of sedatives, narcotics, and anesthetic agents. Phrenic nerve dysfunction can occur after traumatic injuries (e.g., high cervical spine lesions) and is also common after cardiac surgery.14 Diaphragmatic dysfunction may occur following upper abdominal surgery,15 and an elegant study has shown atrophy of diaphragm fibers after only 18 hours of mechanical ventilation and complete diaphragmatic inactivity.16 Critical illness polyneuropathy and myopathy, which are frequent complications of sepsis and multiple organ system failure, may also impede weaning.17,18 Finally, neuromuscular blocking agents (with or without concomitant corticosteroids) and aminoglycosides may contribute to weaning failure.19–24 In addition, malnutrition and deconditioning due to prolonged bed rest/mechanical ventilation can induce severe muscle dysfunction.25
In a multicenter study by De Jonghe et al., a high incidence of intensive care unit (ICU)-acquired neuromuscular dysfunction was reported in patients without preexisting neuromuscular disorders who underwent mechanical ventilation for at least 7 days.18 In this group of 95 patients, 25% were diagnosed with acquired paresis. The duration of mechanical ventilation after the removal of sedation was significantly longer in patients with paresis compared to those who without paresis (18 vs. 8 days; P = 0.03). In this investigation, the independent predictors of ICU-acquired paresis were female sex, number of days with dysfunction of two or more organs, duration of mechanical ventilation before awakening, and administration of corticosteroids. The same group also found that respiratory muscle weakness was associated with delayed extubation.26
Increased Muscle Loads
Increased elastic workloads occur when lung and/or chest wall compliance is reduced (e.g., pulmonary edema, extreme hyperinflation during an acute asthmatic attack, pulmonary fibrosis, abdominal distension, obesity, trauma, or thoracic deformities).13 The presence of intrinsic PEEP is another example of increased elastic workload and is a relatively common phenomenon, especially in patients with COPD.27,28 Dynamic pulmonary hyperinflation, apart from generating an elastic threshold load, places the diaphragm at a mechanically disadvantageous position in which its capacity to generate pressure decreases.
Cardiovascular Dysfunction
The presence of cardiovascular dysfunction can contribute to weaning failure by augmenting loads on the respiratory system and by reducing neuromuscular capacity.29,30 A study by Epstein31 showed that as many as one third of weaning failures resulted solely or in part from congestive heart failure (CHF), although other studies found that fewer episodes of weaning failure (14%) were due to cardiovascular reasons.32 Cardiovascular dysfunction may result from physiologic changes that occur during the resumption of spontaneous unassisted breathing.33 When spontaneous breathing resumes, intrathoracic pressure swings during inspiration are negative, a situation that results in increased left ventricular preload and afterload. A significant decrease in left ventricular ejection fraction has been described during spontaneous breathing trials in COPD patients without coronary artery disease.34
Increased myocardial loading may be sufficient, especially when coupled with left ventricular noncompliance, to precipitate CHF (which stiffens the lungs and further increases respiratory muscle load). Moreover, increased heart loads augment myocardial oxygen demand and may precipitate myocardial ischemia in patients with coronary artery disease.35 Myocardial ischemia causes left ventricular dysfunction that may induce acute pulmonary edema and arterial hypoxemia.
Jubran et al.36 examined hemodynamics and mixed venous saturations in patients during weaning trials. Successfully weaned patients demonstrated increases in cardiac index and oxygen transport compared to values during mechanical ventilation. Patients who failed weaning did not increase oxygen delivery to the tissues owing in part to elevated right- and left-ventricular afterloads. Consequently, these abnormalities can jeopardize respiratory muscle function.
In ICU patients, CHF may be diagnosed for the first time or worsen in patients with this condition as a consequence of increase in venous return, volume overload, or catecholamine release induced by physiologic stress, such as weaning.33,37,38 These factors have negative effects on cardiac function, and together with hypoxemia can result in the development of acute pulmonary edema.33,36,37 Impairment of cardiovascular function can be magnified in the setting of positive fluid balance.39,40
It has been recently shown that performing an SBT in difficult-to-wean patients with a T-tube (instead of pressure support and PEEP) elicits a totally different cardiovascular response and, as expected, as long as support is added (in the form of pressure support and PEEP) the respiratory and cardiovascular function both improve.41
In the ICU there are new noninvasive tools available that help physicians make the diagnosis of cardiovascular dysfunction, such as echocardiography and measurement of plasma B-type natriuretic peptide (BNP).One study found that patients exhibiting weaning failure had higher BNP values than patients who were successfully weaned. Patients who failed weaning were treated with diuretics, and this was accompanied by successful extubation and a decrease in BNP levels.42 Another study compared the use of echocardiography in diagnosing pulmonary edema induced by weaning. The authors showed that an increase in the value of the pulmonary artery occlusion pressure (PAOP) was correlated with echocardiographic signs of increase in left-ventricular filling pressures.43
Mechanisms Explaining Extubation Failure
The reintubation rate may differ according to the etiology of respiratory failure. For instance, in a study that included 217 medical and surgical patients, Vallverdú et al.8 noted that the overall reintubation rate was 15% and ranged from 36% (15 of 42) in neurologic patients to 0% (zero of 13) in COPD patients. The reintubation rate in patients who had acute respiratory failure of other etiologies was 9% (8 of 93). Data by Esteban et al.6,44 indicate that the reintubation rate is about 13% to 19%.
Mechanisms explaining extubation failure include impending abnormalities not diagnosed at the time extubation is performed (e.g., pneumonia, ongoing cardiac failure) and the inability to keep the tracheobronchial tree free of copious secretions.8,45 Intubation can result in laryngotracheal injury, which tends to occur more frequently with increasing duration of intubation and in women, which could explain some episodes of extubation failure.46
Extubation failure results in a marked increase in the duration of mechanical ventilation, ICU and hospital stay, need for tracheostomy, and hospital mortality.6,8,44,47–49 The etiology of extubation failure also influences outcome. Interestingly, patients requiring reintubation because of respiratory failure had a mortality rate of 30%, whereas mortality in patients needing reintubation because of upper airway obstruction was only 7%.49,50 In one study,42 the time to reintubation was found to be an independent predictor of outcome.
Indices to Predict Weaning Outcome
Yang and Tobin51 studied the predictive power of several weaning indices and showed that the rapid, shallow breathing index (f/VT) had the best predictive value. In their study, 95% of patients with a ratio f/VT greater than 105 failed during a test of spontaneous breathing. In general, except for f/VT, these indices exhibit relatively poor positive and negative predictive values. In addition, the performance of these indices is affected by a number of factors, such as selection bias, outcome misclassification, and confounding variables.52
The rapid, shallow breathing index appears to be the most useful bedside method for screening a patient for readiness for liberation. If the value is less than 105, 30 to 120 minutes of an SBT should be used as confirmation of the patient’s capability of breathing spontaneously without assistance. Screening tests are typically performed when the pretest probability of a condition is low. High-sensitivity tests (as is the case with f/VT) are very useful for screening: weaning success is high among patients in whom the test is positive (f/VT <105) and low among those in whom the test is negative (f/VT >105). However, since f/VT has low specificity (a relatively large proportion of weaning-failure subjects in whom the test is positive), the f/VT alone is insufficient to predict weaning failure. For this reason, clinicians utilize additional testing (i.e., 30-120 minutes of SBT).53 From a practical point of view, the information conveyed by weaning indices and clinical judgment should be considered together in making clinically important decisions about extubation.43
Indices to Predict Extubation Failure
The frequency of reintubation and the adverse impact it has on survival indicate that accurate prediction of extubation outcome is important. Most clinicians assess patient readiness for both liberation and extubation by conducting an SBT of variable duration. The crucial importance of performing an SBT before deciding on extubation has been highlighted by Zeggwagh et al.54 These authors proceeded directly to extubation (without performing an SBT) after medical ICU patients had demonstrated clinical improvement. Of the 119 episodes of extubation, 44 (37%) subsequently required reintubation. This rate is much higher than that reported for patients who were extubated after passing an SBT.
Patients incapable of protecting their airway and clearing secretions with an effective cough are at increased risk for extubation failure. Traditional assessment has consisted in demonstrating a cough reflex when the airways are stimulated with a suction catheter and by the absence of excessive secretions, but these criteria have not been standardized. In mechanically ventilated subjects, a “sawtooth” pattern on the flow-volume curve indicates the presence of excess airway secretions but does not provide quantitative information.55
Although tolerance of an SBT up to 120 minutes is a good predictor of successful extubation, Vallverdú et al.8 noted that a high percentage (36%) of neurologic patients who successfully passed a 2-hour SBT and were extubated needed subsequent reintubation. Coplin et al.56
