Respiratory failure exists when the respiratory system cannot maintain gas exchange, causing dysfunction in other organs or threatening life. Such impairment primarily affects oxygenation, manifested by hypoxemia, or affects ventilation, manifested by hypercapnia and respiratory acidosis. This chapter deals with this latter circumstance, commonly called ventilatory failure.
Carbon dioxide (CO 2 ) tension in the arterial blood (Pa co 2 ) is a function of alveolar ventilation ( ) and CO 2 production ( ), according to the following relationship.
Total minute ventilation is the sum of both and dead space ventilation. Either a decrease in total minute ventilation or an increase in dead-space ventilation can thus decrease . Any decrease in or increase in relative to results in an increase in arterial P co 2 . Because bicarbonate retention by the kidney in response to hypercapnia is slow, a sudden increase in arterial P co 2 will not be buffered quickly by bicarbonate and thus will abruptly lower arterial pH. Ventilatory failure exists whenever arterial P co 2 is substantially elevated, and acute ventilatory failure is present when the change from the patient’s baseline state develops rapidly enough to produce a clinically important drop in arterial pH. Because patients with severe chronic obstructive pulmonary disease (COPD), chronic neuromuscular disease, and other disorders may already have hypercapnia at baseline, the presence of a component of acute (acute-on-chronic) ventilatory failure is determined not so much by the arterial P co 2 value as by the presence of acidemia, typically to an arterial pH of less than 7.35. The presence of acute ventilatory failure cannot be determined accurately by physical examination, pulse oximetry, exhaled CO 2 , or other noninvasive tests. Thus, making the diagnosis requires arterial blood gas analysis.
Pathophysiology of Acute Ventilatory Failure
Alveolar ventilation becomes inadequate in relation to CO 2 production either because of a failure of the patient’s ventilatory capability (pump failure) or ventilatory effort (drive failure) ( Fig. 99-1 ). These two mechanisms are distinct in their clinical presentations. Patients with acute failure of the ventilatory pump are dyspneic and tachypneic with other signs of distress, whereas patients with failure of ventilatory drive are not short of breath and typically demonstrate bradypnea or apnea.
Although acute ventilatory failure is primarily a disorder of alveolar ventilation, hypoxemia is also usually present. Alveolar hypoventilation causes a proportional fall in alveolar oxygen pressure (P ao 2 ), according to the alveolar gas equation.
where arterial P co 2 is assumed to be nearly the same as alveolar P co 2 , P io 2 is the inspired P o 2 (i.e., the inspired oxygen fraction multiplied by the difference of barometric pressure and 47 [water vapor pressure at body temperature]), and R is the respiratory exchange ratio. This relationship explains the fall in arterial P o 2 that accompanies alveolar hypoventilation, as shown in Figure 99-2 . Calculating alveolar P o 2 using Equation 2 permits determination of the alveolar-arterial oxygen pressure difference ((A–a)P o 2 ) (commonly but imprecisely called the “A–a gradient” and more precisely called the “A–a oxygen tension difference”). This calculation distinguishes between pure hypoventilation as an explanation for hypoxemia (in which case, (A–a)P o 2 is normal) and the presence of other mechanisms such as low ventilation-perfusion ( ) ratios and right-to-left shunt (in which case, (A–a)P o 2 is increased).
Finally, hypercapnia can be a feature of hypoxemic respiratory failure if the derangement in gas exchange is sufficiently severe. Both right-to-left shunt and low ratios are present in the acute respiratory distress syndrome (ARDS) and can increase dead space–to–tidal volume ratios (V d /V t ) as determined by the Bohr equation, thus impairing CO 2 elimination and contributing to hypercapnia.
Table 99-1 classifies the typical clinical presentations of acute ventilatory failure according to the site or type of defect and the physiologic mechanism or category of disorder responsible. Not every listed example is discussed in this chapter.
More than one mechanism may coexist in a given patient, producing a life-threatening condition even when the individual processes are only moderate in severity. For example, in decompensated obesity-hypoventilation syndrome, a patient whose underlying respiratory drive is reduced and whose obesity poses an increased elastic load on the ventilatory pump may develop acute-on-chronic ventilatory failure in the presence of a relatively modest increase in the work of breathing from an additional restrictive effect of cardiomegaly and pleural effusions.
Acute Ventilatory Failure Due to Insufficient Ventilatory Drive
Congenital disorders that may be associated with diminished hypoxic or hypercapnic ventilatory drive include primary alveolar hypoventilation (or Ondine’s Curse), Prader-Willi syndrome, hypogonadism treated with exogenous testosterone, and Arnold-Chiari malformation. These disorders contribute to the development of acute ventilatory failure through diminished ventilator drive, often in combination with other contributing mechanisms (such as acute infection), and are most often seen in pediatric settings.
Decreased ventilatory drive is a frequent contributor to the development of chronic ventilatory insufficiency but is typically not the sole mechanism for acute ventilatory failure, except for those patients presenting with drug-induced suppression.
Depression of the drive to breathe by drugs is by far the most common circumstance of this form of acute ventilatory failure. The opioids are potent depressors of both hypoxic and hypercapnic ventilatory drive; however, any sedative, hypnotic, or anxiolytic agent causes respiratory depression if administered in sufficient quantity. Propofol, in particular, is a potent respiratory depressant used commonly for sedation during procedures or during mechanical ventilation and must be dosed with caution in patients breathing spontaneously. Respiratory depression resolves as the drug is cleared from the body or is pharmacologically antagonized, as signaled by the return of spontaneous breathing efforts. Because the central nervous system effects of some agents may wax and wane due to the enterohepatic circulation, lipid storage, or other mechanisms, patients should be observed until it is clear that the ventilatory drive has been reestablished before they are weaned from ventilatory support. In the case of drug overdoses and other poisoning, identifying the specific agent or agents involved and the use of any specific therapies available, such as antidotes or dialysis, is important to expedite the weaning process from ventilatory support. Patients who fail to wean from mechanical ventilation because of inadequate drive from depressant drugs breathe slowly or not at all when ventilatory support is briefly discontinued. The much more common reason for difficulty weaning after drug overdose is that other mechanisms (e.g., aspiration pneumonia or sepsis) intervene. In these circumstances, with a return of respiratory drive, patients become tachypneic and manifest signs of respiratory distress when ventilatory support is discontinued.
Other Acquired Causes
Myxedema may present with hypercapnia related to acquired depression of ventilatory drive, and hypothyroidism may be a cofactor when patients present with worsening hypercapnia. Thyroid function testing should be done routinely in patients with new or worsening hypercapnia, especially when a clear physiologic explanation is inapparent. Patients with underlying abnormalities of ventilatory drive who develop respiratory infections, congestive heart failure, or other acute illnesses are more likely to develop acute ventilatory failure compared with individuals without such defects.
The obesity-hypoventilation syndrome, discussed in Chapter 89 , is characterized by blunted responses to hypoxia and hypercapnia and is one disorder in which patients may first present in acute ventilatory failure. Typically, such patients have a history of recent weight gain and are found to be markedly fluid-overloaded, with features of cor pulmonale. The increased work of breathing from decreased chest wall compliance, cardiomegaly, and often large pleural effusions, as well as worsening hypoxemia, all contribute to respiratory muscle fatigue and the development of severe hypercapnia and respiratory acidosis. Hypoventilation and acute ventilatory failure related to obesity are more common among hospitalized patients with severe obesity than has previously been reported. As the prevalence of obesity increases, ventilatory failure from decompensation of the obesity-hypoventilation syndrome is likely to be encountered more frequently.
Acute stroke is another setting in which disordered ventilatory drive may contribute to acute ventilatory failure, even though other processes are usually present, especially the inability to protect the lower airway and to clear respiratory tract secretions. The prognosis of patients with ischemic and hemorrhagic strokes who require intubation and mechanical ventilation is poor both in the short and long term, with a survival rate of 50% at 30 days and 30% after 1 year. Prognosis is especially unfavorable after basilar artery occlusion due not only to depressed respiratory drive but also to the swallowing impairment and problems with secretion clearance seen with brain stem injury.
Although data are not encouraging and the mortality rate is still high, American Heart Association guidelines recommend mechanical ventilation for patients with acute stroke if needed (Level IC). Of course, this necessitates an ethical discussion with the patient and/or family to determine the patient’s wishes regarding aggressive support (see Chapter 104 ).
Principles of Management
Because the underlying physiologic defect is inadequate ventilatory drive despite a presumably normal ventilatory pump, management focuses on restoring normal alveolar ventilation. Although noninvasive ventilation (NIV) is being applied in an increasing number of clinical settings to augment alveolar ventilation, its utility lies mainly in maintaining respiration in patients with failure of the ventilatory pump, whereas endotracheal intubation is generally necessary following acute failure of respiratory drive. NIV can be quite effective in treating chronic congenital or acquired central hypoventilation in outpatients, but in the setting of acute loss of respiratory drive, invasive mechanical ventilation restores alveolar ventilation more rapidly and reliably and is more effective for airway protection and secretion clearance than is NIV.
Because of the patient’s impaired ventilatory drive, a ventilator mode that provides full support, such as volume-targeted assist-control ventilation, should initially be chosen. In the absence of acute lung injury or severe airflow obstruction, ventilator settings should be chosen to aim for values of arterial pH and arterial P co 2 in the normal range, and supplemental oxygen should be supplied (with positive end-expiratory pressure [PEEP] if necessary) to maintain normal arterial P o 2 . Unless there are serious coexisting pulmonary conditions, weaning should be carried out as soon as there is evidence that ventilatory drive is restored (see Chapter 101 ). Extubation is probably safe if the patient has a spontaneous cough, does not require frequent suctioning, and is judged to be able to protect the airway, even if alertness remains impaired.
Acute Ventilatory Failure Due to Neural Transmission Impairment (see Chapter 97 )
Cervical Spinal Cord Injury
Injury to the upper cervical spinal cord may interrupt transmission of the stimulus to breathe from the respiratory centers in the brain stem to the diaphragm and other ventilatory muscles, depending on the injury level. Because the phrenic nerve roots that supply the diaphragm arise from spinal segments C3 to C5, patients with acute injury at this level or above usually require ventilatory assistance. Patients with C1–C2 spinal injury levels are permanently ventilator dependent, whereas those with C3–C4 injuries may eventually achieve at least partial ventilator independence. Lesions below C4 are usually compatible with unassisted ventilation unless there are complicating processes such as intrinsic lung disease or impaired mental status.
Adverse physiologic effects can happen within the first days or weeks after the injury, including loss of lung volumes and inability to take deep breaths (which predisposes to atelectasis), inability to cough normally (which predisposes to the development of pneumonia and complicates its management), and impaired hypoxic pulmonary vasoconstriction (which predisposes to severe and often refractory hypoxemia after atelectasis or pneumonia). These physiologic effects depend on the level of injury, being more frequent with lesions above C4, and on the degree of injury, being more frequent with complete, than with incomplete, lesions.
The short-term prognosis of spinal cord lesions is generally related to the level of the injury, even though some retrospective studies have shown that both mortality and intensive care unit (ICU) length of stay are more strongly influenced by the development of pneumonia and other respiratory complications than by the specific cord injury level.
Although there are reports of initial management of patients with high cervical spinal cord injury (C3-4 or higher) with NIV, great expertise is required to avoid aspiration and other complications. Decisions about NIV should be made on a case-by-case basis ; in most centers, invasive ventilatory support is preferable, at least initially. Phrenic nerve or diaphragm pacing, permitting extubation or removal of the tracheostomy tube (decannulation), has also been reported in the later period following injury. The eventual ability to wean from ventilatory support and to undergo decannulation from tracheostomy are major determinants not only of survival but also of quality of life for patients with cervical spinal cord injury.
Motor Neuron Disease
Amyotrophic lateral sclerosis (ALS) and other motor neuron diseases demonstrate a variable but progressive weakness of the bulbar and ventilatory muscles. This progression determines the course of ventilatory failure and pulmonary complications, which represent the most common cause of death in these patients. Typically, ventilatory muscle weakness develops gradually after the diagnosis is already well established; therefore repeated assessments during outpatient evaluation are useful to monitor the progression of ventilatory muscle impairment. This permits intervention with NIV or, less often, tracheostomy before the onset of acute ventilatory failure. However, some cases present with acute ventilatory failure as the initial manifestation of the disease.
Elective initiation of noninvasive ventilation is becoming a standard of care in motor neuron disease patients with progressive ventilatory impairment because it improves both quality of life and survival rate in patients without significant bulbar involvement. NIV has been successful not only in the chronic, slowly progressive setting, but also in acute ventilatory failure complicating ALS. However, bulbar weakness and a high risk of aspiration make invasive mechanical ventilation a preferred choice for ALS patients with acute ventilator failure, at least initially. With appropriate counseling about end-of-life planning, only a small proportion of patients with ALS receive invasive mechanical ventilation, and presentation to an emergency department with acute ventilatory failure should be unusual.
Injury or Disease Affecting the Phrenic Nerve
Loss of diaphragm function leading to ventilatory failure is often due to spinal cord injury, immunologic diseases (such as Guillain-Barré syndrome or multiple sclerosis), or neuropathy (ALS, Charcot Marie Tooth). Unilateral paralysis of the diaphragm leading to ventilatory failure may be a result of phrenic nerve injury or disease. Its presentation ranges from an incidentally discovered radiographic abnormality without clinical impact to acute ventilatory failure requiring long-term mechanical ventilation, although the latter is quite unusual and generally due to multiple factors. In the past, unilateral diaphragmatic palsy following phrenic nerve injury was most often caused by cold cardioplegia during open-heart surgery or direct injury during internal mammary artery harvesting. Since routine use of insulation for the phrenic nerves during cardiac surgery, this complication is now rare, but unilateral or bilateral phrenic nerve palsies are still seen as a consequence of direct invasion by neoplasm, infectious diseases (such as herpes zoster and Lyme disease), metabolic peripheral neuropathy (diabetes or porphyria), and radiotherapy. Although acute ventilatory failure as a consequence of bilateral diaphragmatic paralysis is unusual, these patients are much more symptomatic than those with unilateral paralysis and usually have severe orthopnea.
Guillain-Barré syndrome, now known as acute idiopathic demyelinating polyneuropathy (AIDP), is an autoimmune polyneuropathy that accounts, together with myasthenia gravis, for the majority of admissions for ventilatory failure due to neuromuscular impairment. Therapy with plasma exchange and intravenous immunoglobulin improves outcomes in AIDP, although 2% to 10% still die, and up to 20% of individuals who survive remain seriously disabled. Theoretically, death should be preventable in the vast majority of patients with this disease because mortality is primarily from potentially avoidable respiratory complications (see the discussion in Chapter 97 ). It remains unclear whether the need for mechanical ventilation can be predicted before the onset of frank ventilatory failure in this condition.
Neuromuscular Junction Impairment
Myasthenia gravis is less common than AIDP as a cause of acute ventilatory failure, although up to 15% to 20% of myasthenic patients experience a crisis during their lifetime. These events usually happen in patients with an established diagnosis of myasthenia gravis. With adequate therapy (plasmapheresis and intravenous immunoglobulin) and respiratory support through noninvasive or invasive mechanical ventilation, the mortality rate is 5% to 10%. Isolated ventilatory muscle weakness requiring mechanical ventilation has been reported as the initial manifestation of the disorder.
Botulism remains an infrequent but important cause of acute ventilatory failure worldwide. In western countries, the incidence of ventilator failure due to food-borne illness has been unusual but steady in recent decades, with about 23 cases/year in the United States. On the other hand, due to subcutaneous injection of black-tar heroin, the incidence of wound botulism has been rising since the 1990s among injectable drug users. The vast majority of patients with both forms of botulism manifest respiratory symptoms and up to 75% of individuals develop clinically significant respiratory failure due to progressive descending flaccid paralysis, requiring mechanical ventilation, which is usually more prolonged in wound botulism patients. Recovery may likewise be prolonged, with residual ventilatory muscle weakness detectable as long as 2 years after presentation.
Primary myopathies due to muscular dystrophies or other congenital myopathies are an uncommon cause of acute respiratory failure in most acute care hospitals but are more common in the pediatric setting. These patients usually develop ventilatory failure gradually and are begun on NIV in an outpatient setting. When they present with acute ventilator failure, it is usually in the setting of a precipitating factor such as pneumonia or bronchitis that causes problems with secretion retention. In such an event, they should be placed in an intensive care unit and treated with an aggressive regimen to assist with secretion clearance. Endotracheal intubation may be necessary to control secretions with weaning to NIV once the acute crisis subsides. Dermatomyositis may also cause respiratory muscle weakness severe enough to lead to acute ventilatory failure, although not as an initial manifestation in the absence of other typical symptoms and signs of this condition. In these reported cases, ventilatory function recovered as the disease was brought under control with immunosuppressive therapy.
Neuromuscular blocking drugs are sometimes administered to ventilated patients, in conjunction with sedation, to facilitate mechanical ventilation, reduce oxygen consumption, or control intracranial pressure. The clinical kinetics of these agents have been determined mainly in the context of short-term general anesthesia, and their effects on ventilatory muscle function in critically ill patients are much more variable. For example, most neuromuscular blocking drugs are cleared more slowly in the presence of hepatic or renal insufficiency. This is particularly true for pancuronium and vecuronium; the effects of these drugs can last days or even weeks in the presence of renal failure. In contrast, atracurium and cisatracurium are metabolized in plasma and do not depend on renal or hepatic function for clearance; thus, they are not associated with prolonged muscle weakness as a result of delayed clearance.
Train-of-four stimulation can be used to monitor the depth of neuromuscular blockade, avoiding excessive paralysis and reducing the quantity of drug used, as well as the recovery time of neuromuscular function in critically ill patients. Although these benefits may not be seen when atracurium and cisatracurium are used, train-of-four testing is sufficiently simple and inexpensive to perform that many experts believe it should be employed routinely. Minimizing the use of neuromuscular blocking drugs in ventilator management and using train-of-four stimulation to monitor the degree of muscle relaxation, as well as employing daily interruptions of paralysis, may reduce the incidence of prolonged paralysis.
Neuromuscular Weakness Associated with Critical Illness
Neuromuscular dysfunction associated with critical illness commonly contributes to the subsequent inability to wean such patients from mechanical ventilation. Several forms of critical illness–associated neuromuscular dysfunction are recognized.
Intensive Care Unit–Acquired Weakness
Unexpected acute weakness and prolonged ventilatory failure were first reported in patients with status asthmaticus treated with corticosteroids and neuromuscular blocking drugs. Subsequently, a similar syndrome was recognized in other groups of ICU patients, especially those with sepsis and systemic inflammation, even without corticosteroid administration or therapeutic paralysis. Either muscle or nerve abnormalities can predominate, leading to a confusing array of diagnostic terms, such as “critical illness myopathy,” “critical illness polyneuropathy,” “postparalytic myopathy,” “ICU-acquired paresis,” “acute quadriplegic myopathy,” and the preferred term, ICU-acquired weakness .
The pathophysiology of ICU-acquired weakness is poorly understood but may involve elements of disuse and active muscle catabolism sparked by systemic inflammation. Electromyography reveals reduced compound muscle action potentials on motor nerve stimulation (with normal conduction velocity); increased action potential duration; and spontaneous electrical activity on muscle needle recording (e.g., fibrillation potentials, positive sharp waves). Biopsy findings may include primary axonal degeneration, type II muscle fiber atrophy, thick filament (myosin) loss, and (occasionally) necrotizing myopathy.
One quarter to one half of all patients who require more than 7 days of ICU care and the majority of patients who develop the systemic inflammatory response syndrome can be shown by neurophysiologic testing to have ICU-acquired weakness. These neurophysiologic abnormalities arise early, accumulate during the course of illness, and usually affect both nerves and muscles. Prospective studies have shown that about one third of critically ill patients exhibit weakness on clinical evaluation. The typical patient exhibits symmetrical extremity weakness in which proximal function is more impaired than distal function and the facial muscles are spared. Clinically, this disorder can produce severe neuromuscular weakness, often affects the respiratory muscles, and may prolong the need for ventilatory support. This syndrome should be suspected in patients who are weak ( Medical Research Council score < 48 ) in the context of critical illness, have the typical clinical examination, and in whom no better alternative cause for weakness can be identified. Handgrip strength may serve as a simple test to identify ICU-acquired weakness. Nerve conduction studies, electromyography, and muscle biopsy generally are not necessary, but their role in diagnosis remains an area of active investigation. The prognosis for recovery of strength is variable, with many patients improving rapidly over days to weeks while others remain weak for many months or longer. The incidence of ICU-acquired weakness may be reduced by intensive insulin therapy, avoiding neuromuscular blocking drugs and corticosteroids where possible, and possibly by early mobilization. Because ICU-acquired weakness is so strongly associated with severity of illness, length of ICU stay, and the presence of multiple organ system dysfunction, prevention focuses on scrupulous attention to good general ICU care and avoidance of sepsis.
Ventilator-Induced Diaphragmatic Dysfunction
Many critically ill patients develop muscle weakness that impedes functional recovery and is associated with prolonged mechanical ventilation. Some component of this represents ICU-acquired weakness, but mechanical ventilation itself (without systemic inflammation) can induce respiratory muscle weakness. The lack of neural stimulation or associated contraction plays a role in the evolution of weakness because measures to keep muscles contracting can ameliorate weakness. The diaphragm, the muscle most responsible for sustaining the work of breathing, may be even more sensitive than other skeletal muscles to the effects of critical illness. In animal models, the diaphragm weakens during the first 1 to 3 days of mechanical ventilation. Using phrenic nerve stimulation, one study demonstrated weakness by measuring a reduction in maximal transdiaphragmatic pressure in a group of continuously ventilated patients in comparison with findings in normal volunteers. As in studies of peripheral skeletal muscle, stimulating the diaphragm attenuates the loss of strength. This suggests that partial rather than full ventilatory support may serve to maintain diaphragmatic strength, potentially reducing time on the ventilator.
Assessment of Need for Mechanical Ventilation in Neuromuscular Weakness
Respiratory muscle weakness can be suspected when there is obvious peripheral muscle weakness, but neuromuscular abnormalities may not always be evident. Respiratory muscle weakness should also be suspected when dyspnea is out of proportion to radiographic and respiratory mechanical abnormalities seen during mechanical ventilation. Orthopnea raises the possibility of diaphragmatic weakness or paralysis. Further, suspicion is raised when maximal inspiratory pressure is reduced or, in some instances, when ultrasound examination of the diaphragms is abnormal.
Early clinical indicators of the need for mechanical ventilation in patients with neuromuscular weakness remain controversial. In addition to subjective assessments of symptoms of dyspnea and respiratory distress, objective assessments of vital capacity and maximum inspiratory and expiratory pressures have been used to evaluate ventilatory muscle capability.
In AIDP, rapid disease progression, bulbar and bilateral facial weakness, and dysautonomia are highly correlated with the need for intubation and mechanical ventilation. Moreover, a reduced vital capacity (<20 mL/kg), maximum inspiratory pressure (less negative than −30 cm H 2 O), and maximum expiratory pressure (<40 cm H 2 O) are associated with the need for intubation. However, no prospective randomized studies have assessed these variables, and predictors of the need for intubation may merely reflect the criteria in current use to determine when a patient should be intubated.
One study on AIDP reported an association between electrophysiologic evidence of demyelination and the need for intubation and mechanical ventilation. Another study on 44 AIDP patients who required mechanical ventilation showed greater cranial nerve involvement and immunoglobulin G (IgG) anti-GQ1b antibody levels than 87 AIDP patients not requiring intubation. These could also be markers of greater disease severity, thus explaining the higher intubation rates.
In myasthenia gravis (MG), the criteria to predict the need for intubation are not as reliable as for AIDP, mostly due to the fluctuating nature of MG. Nonetheless, serial assessments of vital capacity and the use of the same predictors as for AIDP are still recommended in MG as long as the patient is monitored closely in an ICU and caregivers are prepared for emergent intubation if necessary.
In ALS, as previously noted, serial ventilatory muscle assessments during outpatient management are required for the timely initiation of NIV to avoid respiratory crises and the need for emergency intubation. Evaluation of clinical signs of respiratory muscle weakness (such as use of accessory muscles and paradoxical or diminished excursion of the abdomen) and symptoms of diaphragm weakness (such as orthopnea) are important to assess. We recommend serial measurement of pulmonary function tests, nocturnal oximetry, and sniff inspiratory pressure as an index of diaphragm strength. Recommendations for starting NIV in ALS patients vary widely from FVC less than 80% predicted, with the idea that deterioration in respiratory muscle function can be slowed, to less than 50% predicted in the United States as per the threshold for Medicare reimbursement for NIV. The authors believe that data are insufficient to make firm recommendations on any specific FVC threshold for NIV initiation but that, in association with pulmonary dysfunction, NIV should be started when patients develop symptoms that are likely to respond to NIV. For example, dyspnea at rest or orthopnea can respond to NIV, as can symptoms attributable to poor sleep, such as daytime fatigue, hypersomnolence, or morning headaches.
Although the optimal means for monitoring respiratory muscle function remain uncertain, it is clear that initiation of NIV or, if that fails, intubation and mechanical ventilation (if patients desire it) should be undertaken before the development of severe respiratory acidosis or respiratory arrest. For this reason, patients with acute neuromuscular disease who show signs of pulmonary compromise should be monitored in an ICU. Although the rate of progression may fluctuate, serial measurements of vital capacity and maximal inspiratory pressure along with repeated physical examinations focusing on bulbar function and ability to cough are advisable to avoid emergent intubations ( Fig. 99-3 ).
Principles of Ventilator Management
Patients with acute ventilatory failure due to neuromuscular disease usually have normal underlying lung parenchyma. Although NIV is often used successfully in these patients, significant bulbar involvement is associated with a high likelihood of NIV failure and patients with these deficits should be intubated if the goal is prolongation of life. Intubated patients are usually supported with volume-targeted ventilators using tidal volumes of 6 to 8 mL/kg, rates slightly below spontaneous, and 5 to 10 cm H 2 O PEEP to prevent atelectasis. These targets can also be achieved using portable pressure-limited ventilators with the pressure difference between inspiratory and expiratory pressure adjusted to achieve similar tidal volumes (pressure difference usually at least 8 to 10 cm H 2 O) while avoiding respiratory alkalosis. In patients requiring continuous ventilatory assistance, settings should aim not only to maintain gas exchange but also comfort, considering that there is no convincing evidence that “exercising” respiratory muscles in neuromuscular disease speeds recovery. Some advocate a new protocol using prophylactic NIV and mechanical cough assistance to facilitate extubation and avert the need for reintubation in patients with neuromuscular disease who are weaning from invasive mechanical ventilation.
Acute Ventilatory Failure Due to Chest Wall Defects (see Chapter 98 )
Many restrictive diseases of the lungs or chest wall progress insidiously over months or years. Critical illness may represent the natural history of the underlying condition but may also signal an acutely superimposed, potentially reversible crisis such as infection, pneumothorax, or thromboembolism.
Chest Wall Skeletal Abnormalities
Thoracic restriction and ventilatory muscle dysfunction due to severe kyphoscoliosis typically lead to gradually progressive ventilatory insufficiency. Such patients can present with acute or acute-on-chronic ventilatory failure and require intensive care. Physiologic studies have shown that both lung and chest wall mechanics are impaired during acute respiratory failure in patients with kyphoscoliosis. Chest trauma, especially when leading to flail chest due to rib fractures, may also cause the development of acute hypercapnic respiratory failure. In either case, the ventilator pump fails because of inability to sustain ventilatory work due to abnormalities that compromise ventilatory function, such as increased chest wall stiffness in kyphoscoliosis and decreased ventilator efficiency due to paradoxical chest wall motion and pain in flail chest.
Primary disease of the pleura, such as asbestos-related diffuse pleural thickening or postinflammatory fibrothorax, could potentially present in a similar way as skeletal deformities, but dyspnea and hyperventilation are more common with these chronic pleural diseases. Respiratory acidosis develops late in the course of the disorder unless ventilatory drive is depressed or there is concomitant lung involvement. Pleural effusion or pneumothorax can likewise precipitate acute ventilatory failure, usually in the presence of underlying obstructive or restrictive pulmonary parenchymal disease.
Principles of Management
Long-term NIV appears to be beneficial in selected patients with kyphoscoliosis and other chest wall diseases and has been reported to be successful in acute-on-chronic ventilatory failure. Some recent studies report that NIV reduces the need for intubation and leads to a shortened hospital stay in patients with chest trauma.
Parenchymal Lung Disease
Idiopathic pulmonary fibrosis and other pulmonary parenchymal restrictive diseases are usually associated with hyperventilation rather than hypoventilation. However, acute ventilatory failure can arise in the late stages of these conditions, either as a manifestation of the primary disease process or, more often, in conjunction with pneumonia, surgery, or other intercurrent illness. Physiologic assessment has demonstrated marked increases in lung stiffness and airway resistance in patients with end-stage idiopathic pulmonary fibrosis requiring mechanical ventilation, explaining the development of hypercapnia and acute ventilatory failure.
Several case series have documented the poor prognosis of patients who present with acute respiratory failure in the setting of advanced interstitial fibrosis. In one retrospective report, all 14 consecutive patients with acute respiratory failure and idiopathic pulmonary fibrosis admitted to the ICU died despite aggressive ventilatory support. In another report of 23 similar patients, 22 patients died; the single survivor received a lung transplant shortly after admission. In a third series of 19 patients with idiopathic pulmonary fibrosis and ARF, 13 died. Outcomes appear equally poor regardless of whether ventilation is invasive or noninvasive.
Principles of Ventilator Management
Due to increased lung stiffness, NIV for restrictive disease usually requires higher airway pressures than are used for COPD. Thus, avoiding gastric insufflation and air leaks around the mask is more challenging. Furthermore, considering that a superimposed condition such as a respiratory infection often precipitates the acute bout of respiratory failure, accompanied by increased work of breathing and secretion retention, invasive ventilatory support is often warranted. The best way to ventilate patients with acute respiratory failure in the setting of underlying restrictive thoracic or lung disease has not been determined by clinical trials. The potential for hemodynamic compromise, barotrauma, and ventilator-induced lung injury with the use of high pressures and the physiologic similarity to ARDS in patients with deteriorating pulmonary fibrosis make it reasonable to apply similar lung-protective ventilator strategies and management targets (see Chapter 101 ). Low tidal volumes (e.g., 6 mL/kg predicted body weight) should be applied, attempting to keep the end-inspiratory plateau pressure below 30 cm H 2 O if possible. Patients with restrictive lung disease typically breathe rapidly and shallowly, so tachypnea may not be avoidable during the weaning process and should not be used as the sole reason for delaying extubation if gas exchange and other assessments are acceptable.
Acute Ventilatory Failure Due to Airway Obstruction
Upper Airway Obstruction
Upper airway obstruction is an occasional cause of acute ventilatory failure. The onset can be precipitous, as with occlusion of the glottis by an aspirated foreign body (i.e., “café coronary”) or a swollen and edematous epiglottis due to acute epiglottitis. The onset can also be insidious, progressing over months, as with a tracheal tumor. The severity and length of narrowing and air flow determine the airway resistance and thus the additional work of breathing imposed by the obstruction. Gradually progressive upper airway narrowing may be well tolerated, at least while breathing at rest, until a critical limit is reached, often when the airway diameter drops to the range of 5 to 6 mm.
The location and variability of the narrowing are also important in determining the clinical manifestations. Extrathoracic variable upper airway narrowing affects mainly inspiratory flow because the negative intraluminal pressure exacerbates the narrowing during inspiration. During expiration, the positive intraluminal pressure widens extrathoracic airways. Vocal cord paralysis is an excellent example of a variable extrathoracic upper airway obstruction, producing stridor and severe airway obstruction during inspiration but no significant obstruction during expiration. The opposite pertains to variable intrathoracic obstructions, with narrowing becoming less severe during inspiration because pressure gradients favor airway widening. During expiration, the airways narrow and the severity of the obstruction worsens. Tracheomalacia can cause variable intrathoracic airway obstruction. Fixed obstructions affect both inspiration and expiration regardless of their location.
Upper airway obstruction causes ventilatory failure by increasing airway resistance to the point where respiratory muscles can no longer sustain minute volume at a level adequate to maintain CO 2 homeostasis. Negative-pressure pulmonary edema can also contribute to the gas-exchange impairment. Ideally, the therapeutic aim is to relieve the obstruction. This may be achieved by removal of a foreign body, laser therapy of an endotracheal tumor, placement of a stent in an area of tracheomalacia or stenosis, or tracheostomy to bypass an area of obstruction. Inhalation of heliox (to reduce airway resistance), continuous positive airway pressure (CPAP), or noninvasive ventilation using pressure support and PEEP can help to reduce the work of breathing and avoid intubation in patients who have reversible causes of their upper airway obstruction, such as postextubation stridor, or who are awaiting tracheostomy or surgical repair of an obstruction. However, these temporizing measures require close monitoring with the recognition that the patient can deteriorate abruptly.
Chronic Obstructive Pulmonary Disease
COPD is the third leading cause of death among adults aged 65 to 84 in the United States, being the primary contributor to mortality caused by lower respiratory disease, and poses enormous costs to the U.S. health care system, mostly due to hospitalization. The vast majority of hospitalizations are due to acute exacerbations of COPD, but other causes such as acute pneumonia, congestive heart failure, pulmonary embolism, and pneumothorax may contribute to the deterioration.
Hyperinflation associated with severe COPD places the respiratory muscles at a mechanical disadvantage ( Fig. 99-4 ). The loss of elastic structures is responsible for an increase in lung compliance leading to hyperinflation (with an increase in total lung capacity and functional residual capacity) and the collapse of small airways during expiration that contributes to an increase of residual volume, often referred to as “air trapping.” The flattening of the diaphragm increases the radius of curvature, which, according to the Law of Laplace, also increases muscle tension and impedance to blood flow. In addition, ventilatory efficiency is reduced because the shortened diaphragm operates at a disadvantageous position on its length-tension curve and the horizontal orientation of the flattened diaphragm causes the lower rib cage to move paradoxically during inhalation, inward rather than outward (“Hoover sign”).
The hyperinflation and impairment in diaphragm function necessitate the recruitment of accessory muscles to maintain ventilation at higher lung volumes, contributing to the already increased oxygen cost of breathing. Finally, the collapse of small airways predisposes to incomplete emptying and positive intrathoracic pressure at end-expiration (intrinsic or auto-PEEP). Auto-PEEP poses an inspiratory threshold load requiring that inspiratory muscles lower the elevated alveolar pressure to subatmospheric in order to initiate airflow for the next breath.
During an exacerbation of COPD, the combination of airway swelling, secretions, and bronchospasm caused by acute inflammation increases airway resistance, further worsening the expiratory flow-limitation and increasing end-expiratory lung volume. As depicted in Figure 99-5 , COPD patients adapt by attempting to maintain airflow by breathing at even higher lung volumes. In addition, they adopt a rapid, shallow breathing pattern that further limits the time available for expiration, aggravating intrinsic PEEP and adding further to the work of breathing. The diaphragm flattens more and develops increased tension, further impeding diaphragmatic blood flow. The resulting limitation in substrate delivery to muscle is aggravated by progressive hypoxemia, caused by worsening hypoventilation and imbalance related to secretion retention. Thus, as the demand for breathing increases, the capacity to supply breathing work diminishes. As respiratory drive increases in a futile attempt to reverse the worsening alveolar hypoventilation, muscular performance deteriorates and the diaphragm fatigues. A vicious cycle ensues, leading inexorably to worsening respiratory muscle fatigue, ventilatory failure, and death unless therapeutic interventions interrupt the cycle.
Patients with exacerbations of COPD must be carefully evaluated to identify those at risk of developing respiratory failure and to exclude other causes of respiratory failure. History and physical examination are useful. Although the Borg or visual analogue scales help gauge the level of dyspnea in clinical studies, a subjective assessment that dyspnea is worse than at baseline and of at least moderate severity suffices to identify patients who may be at risk for respiratory failure. Physical findings seen with severe exacerbations include tachypnea; accessory muscle use; abdominal paradox; Hoover sign (inspiratory inward motion of the lower, lateral rib cage); cyanosis; and mental status alterations.
In addition to a sputum examination for purulence, a white blood cell count, electrocardiogram, chest radiograph, and arterial blood gas should be obtained to assess the severity of an exacerbation. The widespread use of continuous pulse oximetry and of venous blood gases has decreased, but not eliminated, the need for arterial blood gases. Whereas venous pH values generally agree with arterial values, venous P co 2 poorly reflects arterial P co 2 ; nonetheless, a normal venous P co 2 may be useful in excluding hypercapnia. Arterial blood gases provide a rapid assessment of arterial P co 2 and pH, information that is critical when deciding to place patients in critical care units or to initiate mechanical ventilation and to assess response to therapy. During severe exacerbations, patients with chronic CO 2 retention develop acute-on-chronic hypercapnia, manifested by a drop in pH indicative of retained CO 2 uncompensated by bicarbonate, an important indicator of ventilatory failure that can be detected only by measurement of arterial blood gases.
Medical therapy, consisting of bronchodilators, corticosteroids, and antibiotics, should be promptly started in patients with severe exacerbations. Additional therapies, including diuretics, nitrates, or anticoagulation, should be started whenever comorbidities such as congestive heart failure or pulmonary embolism are suspected.
Oxygen should be supplemented routinely to improve hypoxemia, but it should be carefully titrated in patients with CO 2 retention to maintain a target Sp o 2 of 88% to 92%. Overzealous oxygen supplementation in such patients has long been known to aggravate CO 2 retention, by either blunting the hypoxic ventilatory drive, increasing physiologic dead space (perhaps due to oxygen-induced bronchodilation in poorly perfused lung regions), or both. Because hypoxemia in COPD patients is usually due mainly to hypoventilation and is easily reversed, initial supplementation with nasal oxygen at 2 L/min is often adequate. In patients with severe exacerbations, arterial blood gases should be repeated periodically to assess the effect of oxygen supplementation on arterial P co 2 .
Although medical therapy alone is usually effective in mild COPD exacerbation, it is often not sufficient in severe exacerbations. In severe exacerbations, tachypnea, dyspnea, and CO 2 retention may persist or worsen despite initial medical therapy. Before 10 years ago, patients in such a predicament would usually be intubated and mechanically ventilated. If they declined intubation, they were kept comfortable while medical therapy was continued, but they often died. Invasive mechanical ventilation was successful in the majority of cases, but hospital mortality rates were substantial, averaging 30% in several studies. Complications of invasive mechanical ventilation were common, including upper airway trauma, pneumothorax, and nosocomial infection, all contributing to patient mortality.
In 1990 Brochard and coworkers demonstrated that the noninvasive delivery of pressurized air into the lungs via a face mask was effective in providing partial ventilatory assistance during COPD exacerbations. These workers used a device designed to provide pressure support that reduced diaphragmatic work of breathing by increasing airway pressure with each inhalation. Later, Appendini and colleagues demonstrated that combining extrinsic PEEP (to counterbalance the effects of intrinsic PEEP) with pressure support even more effectively reduced the work of breathing in COPD patients than either CPAP or pressure support alone. By reducing the work of breathing, NIV restores the balance between supply and demand for the work of breathing, thereby serving as a “crutch” during COPD exacerbations and halting the progression of respiratory muscle fatigue while medical therapies are given time to work.
Since Brochard’s groundbreaking study, multiple randomized controlled studies and meta-analyses have demonstrated the efficacy of NIV to treat exacerbations of COPD. When compared with conventional therapy alone, NIV for severe exacerbations of COPD more rapidly improves dyspnea, respiratory and heart rates, arterial P co 2 , and encephalopathy scores. In addition, intubation and mortality rates drop precipitously (from roughly 75% and 30% in controls to 25% and 10%, respectively, in NIV-treated patients). NIV also lowers complication rates and hospital lengths of stay compared with controls. One study has reported that NIV failed to lower intubation or mortality rates or hospital lengths of stay in patients with COPD exacerbations, but it is notable that blood gases were only mildly deranged and there were no intubations or mortality in the control group. This result suggests that patients with relatively mild COPD exacerbations are unlikely to derive benefit from NIV, and the modality should usually be reserved for those with mild to severe symptoms.
Several meta-analyses have concluded that NIV is effective in avoiding intubation (relative risk 0.42 and absolute risk reduction 28%, respectively), reducing mortality (relative risk 0.41 and absolute risk reduction 10%, respectively), and shortening hospital length of stay (by ≈4 days). A recent study on a large cohort of patients (25,628) with COPD exacerbations requiring mechanical ventilation showed reduced mortality, length of stay, and cost with NIV compared to invasive ventilation. On the basis of this evidence, the authors of these meta-analyses, reviews, and guidelines have advised that NIV should be the ventilatory modality of first choice and should be started early in the course of moderate to severe COPD exacerbations.
Heliox Combined with Noninvasive Ventilation.
By virtue of its lower density than nitrogen, helium can be combined with oxygen to lower the airway resistance attributable to turbulent flow. The oxygen concentration can be increased to about 40% in the helium-oxygen mixture but not higher without losing the density advantage of the added helium. Heliox has been combined with NIV to treat patients with COPD exacerbations, with beneficial physiologic responses including reduced airway resistance and more rapid improvements in gas exchange. However, subsequent randomized, prospective trials on patients with COPD in respiratory failure found that the addition of heliox to NIV offered no significant advantages over NIV alone in terms of intubation or mortality rates or hospital lengths of stay.
COPD Complicated by Pneumonia
COPD patients may develop acute or acute-on-chronic respiratory failure due to an exacerbation complicated by pneumonia. When evaluating a COPD patient with worsening of baseline symptoms, concomitant pneumonia should be considered as a contributing factor. By virtue of impairment in cellular and molecular defense mechanisms and the common use of inhaled corticosteroids, which have been associated with increased pneumonia rates, COPD patients are at risk for pneumonia. In addition, pneumonia is related to a more severe presentation of community-acquired pneumonia in hospitalized patients, without being a risk factor for mortality.
Pneumonia has been associated with a poor outcome in patients treated with NIV. However, in one trial on severe community-acquired pneumonia, NIV reduced intubation (21% vs. 50%, P = 0.03) and mortality rates, and shortened ICU length of stay (1.8 vs. 6.0 days, P = 0.04) compared with standard oxygen therapy. Benefit, however, was confined to the subgroup of patients with underlying COPD. Thus, although the presence of pneumonia is a risk factor for poorer outcome with NIV, COPD patients with pneumonia can still benefit.
Postoperative pulmonary complications are defined as pulmonary abnormalities (such as atelectasis, pulmonary embolism, ALI/ARDS) that arise frequently in the postoperative period, particularly in COPD patients. These complications, due to general anesthesia, postoperative immobility, or to the surgery itself, increase morbidity, mortality, and length of stay.
NIV has been reported to be effective in reducing the need for intubation, ICU length of stay, and mortality rate in post–lung resection patients with acute respiratory insufficiency. Although only a portion of these patients had COPD, accumulating evidence now supports the use of NIV in selected postoperative (including COPD) patients to maintain improved gas exchange and avoid reintubation and its attendant complications. NIV techniques are also being used prophylactically to reduce secretion problems, atelectasis, and hypoxemia after thoraco-abdominal and major abdominal surgery.
Postextubation in COPD
Between 10% and 15% of patients develop respiratory failure after a standard extubation, increasing the length of stay on mechanical ventilation and in the ICU and therefore the risk of related complications including mortality. In this context, NIV can be used in several ways: (1) to permit earlier removal of the endotracheal tube by assisting ventilation postextubation, (2) to prevent the onset of respiratory failure and need for reintubation in patients at risk for respiratory failure postextubation, and (3) to avoid the need for reintubation in patients who develop frank respiratory failure postextubation.
Using NIV to allow earlier removal of the endotracheal tube is supported by randomized controlled trials. One trial demonstrated that extubation to NIV after 48 hours of intubation increased overall weaning rate after 60 days (88% vs. 68%), shortened the duration of mechanical ventilation (10.2 vs. 16.6 days), shortened the stay in the ICU (15 vs. 24 days), and improved 60-day survival (92% vs. 72%) (all P < 0.05) compared with patients left intubated. A second randomized, controlled trial in patients with “persistent weaning failure” (failure of spontaneous weaning trials on 3 consecutive days) showed that early extubation to NIV significantly reduced ICU and hospital length of stay, incidence of nosocomial pneumonia (from 59% to 24%, P < 0.05), complication rate, and hospital and 90-day mortality (odds ratio 3.5).
These randomized studies support use of NIV to facilitate early extubation of invasively ventilated COPD patients. However, if early extubation is contemplated, it should be reserved for carefully selected patients. Patients should be recovering from COPD exacerbations, be on 15 cm H 2 O or less of pressure support, be able to sustain 5 to 10 minutes of unassisted breathing, have an adequate cough without excessive secretions, be easy to intubate, and have few if any comorbidities.
Using NIV in patients who develop respiratory failure postextubation to avoid reintubation has less support in the literature. Two randomized trials of patients at high risk for extubation failure used NIV prophylactically to prevent reintubation but failed to show the anticipated benefit. In one, NIV provided no reduction in the need for intubation, duration of mechanical ventilation, length of hospital stay, or mortality. In the other, NIV failed to show benefit in these variables and was associated with increased ICU mortality. In the latter study, the increased mortality was thought to be related to a 10-hour longer delay before proceeding with reintubation compared with controls. Furthermore, only 10% of patients in both of these studies had COPD, leading to the speculation that results might have been favorable if more patients with COPD had been enrolled.
This speculation has been borne out by two subsequent randomized, controlled trials, one showing dramatic reductions in respiratory failure, need for reintubation, and mortality in a subgroup of hypercapnic patients, and the other demonstrating that patients with hypercapnia postextubation have a significant reduction in acute ventilatory failure if treated prophylactically with NIV compared with standard oxygen supplementation. Thus, the best current recommendation is to use NIV selectively in patients with extubation failure, mainly for COPD or other hypercapnic patients, and to avoid delays in intubation should NIV fail.
The use of NIV to treat respiratory failure in patients who have declined intubation accounted for 10% of acute applications in one survey. This application has been controversial, with some arguing that there is little to lose because it may reverse the acute deterioration or, at least, provide relief of dyspnea and a few extra hours to finalize affairs. Others have argued that this merely prolongs the dying process, consumes resources inappropriately, and may add to discomfort or counter patients’ wishes about avoiding life-prolonging measures. In prospective observational studies of 113 and 131 do-not-intubate (DNI) patients treated with NIV, survival to hospital discharge was greater than 50% for COPD and congestive heart failure patients, whereas it was lower (14% to 25%) for those with a diagnosis of hypoxemic respiratory failure (pneumonia) or advanced cancer. Thus, NIV can be used to treat respiratory failure for DNI patients with acutely reversible processes such as COPD exacerbations. Alternatively, it can be used to palliate DNI patients, by alleviating dyspnea or providing temporary support. The patient or family should be informed that NIV is being used as a form of life support that may be uncomfortable and can be removed at any time.
Practical Application of Noninvasive Ventilation
A thorough discussion of the application of NIV is beyond the scope of this chapter, and the reader is referred to Chapter 102 and elsewhere for more complete descriptions. The following sections focus on aspects relevant to applications in COPD patients with acute respiratory failure.
Selection of appropriate patients is key to the successful application of NIV. The selection process should take into account the patient’s clinical characteristics and risk of failure on NIV ( Table 99-2 ). Predictors of success of NIV have been identified ( Table 99-3 ) and include a good neurologic status (and hence more cooperativeness), ability to protect the airway, and only mild-moderate acid-base or gas-exchange derangement. Several studies have also found that improvements in pH, arterial P co 2 , and level of consciousness within the first hour or two of NIV initiation are strong predictors of success. These studies indicate that there is a “window of opportunity” when initiating NIV that opens when patients need ventilatory assistance but closes if they progress too far and become severely acidemic. Ultimately, it becomes a clinical judgment that takes into account the patient’s diagnosis that led to the respiratory failure, the need for ventilator assistance, and the absence of contraindications (see Table 99-2 ).