Pulmonary Aspects of Neurological Intensive Care

Pulmonary disorders are a central feature of the practice of intensive care medicine. Furthermore, they are particularly pertinent because certain types of respiratory failure derive primarily from disorders of the nervous system. Additionally, as in other intensive care unit (ICU) settings, many patients with serious neurological conditions have coexisting pulmonary conditions such as asthma or chronic obstructive pulmonary disease (COPD); others develop complications of treatment or immobilization such as ventilator-associated pneumonia (Chapter 7) or pulmonary embolism. However, among the most interesting aspects of the interaction between the nervous and respiratory systems are the conditions summarized by the term neurogenic respiratory failure (NRF), encompassing both central and peripheral nervous causes of pulmonary distress; it is this subject which comprises the bulk of discussion in this chapter. These forms of respiratory failure derive from neuromuscular diseases (Guillain-Barré syndrome, myasthenia gravis, etc.), primary brainstem lesions, and from diseases that cause coma and secondary reduction in airway reflexes and ventilation, all of which are detailed in this volume.

Several general aspects of the subject that are necessary to the intelligent application of ventilators and endotracheal tubes in any ICU are also presented here because they are referable to all types of respiratory failure. Much of this information, particularly classic respiratory physiology which has been considered the core knowledge of respiratory intensive care, should be known to all competent neurointensivists who plan to deal with and discuss intelligently general intensive care problems.

ACUTE RESPIRATORY FAILURE

Respiratory failure may be divided into four broad classes (Table 4.1) (1). Although the first two types are the ones emphasized in most discussions of neurological critical care, the other two should be kept in mind in that there are many patients who benefit from mechanical ventilation but do not fit strictly in one of the first two categories. For example, patients in cardiogenic shock benefit from mechanical ventilation in part because they no longer need to deliver as much oxygenated blood to the diaphragm; this decrease in oxygen demand allows for the diminished cardiac output to be redistributed to other organs. Or, hyperventilation for the short-term control of increased intracranial pressure (ICP) is similar to perioperative mechanical ventilation insofar as one employs mechanical ventilation to meet a therapeutic goal that is not strictly mandated by the patient’s disease.

The unifying feature of the group with failure of oxygenation is an inability to transfer sufficient oxygen from the airways into the pulmonary venous blood. In neurocritical care, most cases of primary oxygenation failure stem from pulmonary edema, pneumonia, pulmonary embolism, or severe atelectasis.

TABLE 4.1. Classification of respiratory failure

	Description	Conventional example	Neurological example
Type I	Oxygenation failure	Cardiogenic pulmonary edema	Neurogenic pulmonary edema
Type II	Ventilation failure	Narcotic overdose	Guillain-Barré syndrome
Type III	Perioperative	Coronary artery bypass graft	Craniectomy for tumor
Type IV	Shock	Cardiogenic shock	Acute spinal cord injury

Pulmonary Edema

Pulmonary edema in all its forms indicates an excess of extravascular lung water. Under normal circumstances, an increase in the transudation of water from the pulmonary capillaries into the interstitial space is handled by the lymphatic system, which is able to increase its flow rate severalfold in order to keep excess fluid from accumulating. The interstitial space enlarges if the rate of fluid transudation exceeds the lymphatic capacity, thereby inhibiting gas exchange by increasing the distance through which gases must diffuse in order to be transferred both to the capillary (oxygen) and from the alveolus (carbon dioxide), the capillary, and the alveolus. The lower solubility of oxygen in water also results in slower exchange, such that even under normal circumstances oxygen does not completely equilibrate between the alveolus and the capillary during the transit time of a red blood cell. As a practical matter, carbon dioxide normally equilibrates completely in less than this transit time; this is why an increase in interstitial lung water produces hypoxemia but rarely hypercarbia.

Fluid transudation further impairs gas exchange by decreasing the number of respiratory units (the parts of the lung capable of gas exchange, including the alveoli, respiratory bronchioles, and terminal bronchioles) available for ventilation. If perfusion of these alveoli continues, the blood leaving these units is still venous in its gas composition; when it mixes with blood from functioning units, the result is a mixture of blood with a low PpvO₂ (venous O₂); that is, a true “shunt” and a slight increase in PpvCO₂ (venous CO₂). The brainstem senses this hypoxemia, however, and increases minute ventilation in response; thus, the measured PaCO₂ is generally lower than 40 mm Hg.

The commonly recognized causes of pulmonary edema can be predicted from the Starling equation

where: Q = the net fluid flow across the capillary membrane; K = the reflection coefficient; P_c = the hydrostatic pressure within the capillary; P_i = the hydrostatic pressure in the interstitium; σ = the permeability constant; π_c = the colloid oncotic pressure in the capillary; and π_I = the colloid oncotic pressure in the interstitium. Thus, conditions that raise capillary hydrostatic pressure (e.g., elevated left atrial pressure in mitral stenosis or congestive heart failure) increase the driving force for fluid to leave the capillary and enter the interstitium. Markedly lowering the plasma oncotic pressure has the same effect but this is a rare clinical occurrence. If the permeability of the capillary to plasma proteins increases (becomes “leaky”), fluid moves out of the capillary and is difficult to return to the circulation until normal permeability is restored.

A further result of increased interstitial lung fluid flux is a decrease in lung compliance (unit change in volume per unit change in airway pressure, or ΔV/ΔP, in mL/cm H₂O, as described in the discussion of pulmonary physiology). For the spontaneously breathing patient, diminished compliance increases work of breathing. For the mechanically ventilated patient, diminished compliance increases the airway pressure required for ventilation and thereby the risk of barotrauma.

Noncardiogenic and Neurogenic Pulmonary Edema

Many patients with neurological disease have coincident cardiovascular diseases that
may produce congestive heart failure, or cardiogenic pulmonary edema, and also may have aspiration or trauma producing acute lung injury (termed noncardiogenic pulmonary edema, such as neurogenic pulmonary edema). One useful classification of acute lung injuries is listed in Table 4.2 (2). In neurocritical care practice, the intensivist is particularly alert to the development of neurogenic pulmonary edema (NPE). The conditions associated with NPE and the presumed physiology of this process are discussed in detail in Chapter 5.

TABLE 4.2. Criteria for the diagnosis of acute lung injury

Item			Value	Score
Consolidation on chest radiograph			None	0
			One quadrant	1
			Two quadrants	2
			Three quadrants	3
			Four quadrants	4
Hypoxemia score (PaO₂:FiO₂ ratio)			>300	0
			255-299	1
			175-254	2
			100-174	3
			>100	4
Positive end-expiratory pressure (PEEP) score			≤5 cmH₂O	0
			6-8 cmH₂O	1
			9-11 cmH₂O	2
			12-14 cmH₂O	3
			≥15 cmH₂O	4
	Final score = Divide aggregate sum by number of components used
		0	No lung injury
		0.1-2.5	Mild-moderate lung injury
		>2.5	Severe lung injury
From Murray JF, Matthay MA, Luce JM, et al. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988;138:720-723, with permission.

Diagnosis

The suspicion of pulmonary edema arises whenever oxygenation becomes abnormal; this may manifest as tachypnea or dyspnea in patients who can make these responses, and decreased oxygen saturation or PaO₂ (or the need to increase the FiO₂ to prevent such decreases). Rales (crackles) reflect alveolar edema, and a third heart sound reflects elevated ventricular end-diastolic volume; these findings have high positive predictive values but relatively low sensitivities (3). When substantial alveolar edema occurs, the patient may produce pink sputum. Other physical findings suggesting congestive heart failure should be sought. However, the clinical suspicion of pulmonary edema should prompt a chest radiograph, because the differential diagnosis of oxygenation difficulty is broad, and the chest film often reveals lung findings other than or in addition to pulmonary edema that require management. It also allows one to estimate the importance of cardiac dysfunction as a contributor to pulmonary edema if present. Echocardiography is a useful adjunct if cardiac dysfunction is suspected. Electrocardiography and measurement of troponin concentrations should be undertaken if there is reason to suspect myocardial damage as the cause of pulmonary edema.

Perhaps even more crucially than others, the patient with acute neurological disease who becomes hypoxemic requires immediate correction of this problem. Patients who have suffered head trauma or have other reasons for cerebral ischemia do not tolerate superimposed systemic hypoxia. Although recent
work indicates that excessive supplemental oxygen may be deleterious to stroke patients (4), one should not fear a brief period of elevated PaO₂ during the prevention or correction of hypoxemia. Furthermore, fear of oxygen toxicity induced by prolonged inspiration of gas at an FiO₂ exceeding 0.50 to 0.60 should never be used as a reason to permit hypoxia.

Patients who are breathing spontaneously may respond to increased oxygen by nasal cannulae or face mask; if continuous positive airway pressure (CPAP) is needed to prevent respiratory units from collapsing, this can be delivered by face mask in some patients. However, many patients with acute neurological disease are unable to cooperate with this treatment. Continuous positive airway pressure masks also impair access to the patient’s airway for suctioning, and may result in massive aspiration if the patient vomits without someone in the room to remove the mask immediately. Thus, most patients with critical illnesses of the nervous system who develop pulmonary edema require endotracheal intubation.

Intubation of these patients often requires some sedation, and may also require neuromuscular junction blockade (as discussed later). Sedative agents have varying effects on intracranial pressure and systemic arterial pressure. Thiopental provides good sedation and coincident decrease in intracranial pressure if the patient has stable autonomic function. We prefer the use of etomidate, which provides brief anesthesia without measurably lowering blood pressure if hypotension is present or considered likely. Should the patient already be comatose and not require sedation, intravenous (i.v.) lidocaine may blunt the increase in ICP associated with intubation and should be given shortly before the procedure. Intravenous lidocaine also should be considered if an awake intubation is planned with topical anesthesia, because the latter does not prevent ICP elevation (5).

If neuromuscular blockade is deemed necessary, one should consider the risk of hyperkalemia produced by a depolarizing agent (e.g., succinylcholine) in patients with neurological or neuromuscular disease. Succinylcholine may also increase ICP unless the patient has been pretreated with a small (“defasciculating”) dose of a nondepolarizing agent. Therefore, in most circumstances, one should rely on a nondepolarizing agent (e.g., vecuronium).

In almost all circumstances, endotracheal tubes should be placed through the mouth rather than the nose. This technique helps avoid nosocomial sinusitis, which is a risk factor in the development of nosocomial pneumonia. Nasal intubation, although mechanically more stable in the long run for awake and moving patients, is more prone to local bleeding if anticoagulation or plasma exchange become necessary. Endotracheal tubes capable of continuous aspiration of subglottic secretions should be considered (see the following).

Although the usual management of cardiogenic pulmonary edema involves diuretic and vasodilator therapy to decrease the hydrostatic pressure driving edema production, this process carries substantial risk in patients with subarachnoid hemorrhage (SAH) and others in whom cerebral ischemia may be present, even if hypotension is not produced. In addition, because NPE can occur at pulmonary capillary wedge pressure (PCWP) substantially lower than those present in cardiogenic edema, attempts to further lower PCWP may produce hypotension. Thus, the initial strategy should concentrate on supplemental oxygen and either CPAP (for the spontaneously breathing patient) or positive end-expiratory pressure (PEEP) (for the mechanically ventilated patient). After initial stabilization, determination of the wedge pressure by direct measurement through a pulmonary artery catheter should be used to guide therapy. A wedge pressure of 15 to 18 mm Hg is the usual goal in this setting. At higher wedge pressures, the patient may be at risk for an additional hydrostatic contribution to edema formation. (A cardiogenic source should be sought if the pressure is elevated at the time pulmonary edema is documented.) Lower wedge pressures may be associated with hemodynamic aggravation of vasospasm in SAH patients.

Positive end-expiratory pressure and CPAP raise intrathoracic pressure and thereby decrease cardiac preload. Thus, one must be alert for their potential effects on blood pressure, particularly in dehydrated patients. In the patient who is already in a state of increased peripheral oxygen extraction (as may occur in hypoxemia), a decrease in preload that adversely affects cardiac output further lowers the oxygen content of venous blood, resulting in incomplete oxygen uptake by the blood during its passage through the lungs, and hence progressive arterial hypoxemia. It is important to transiently decrease airway pressure to determine whether this problem is present, and to increase preload by fluid administration (even in the face of pulmonary edema) if indicated.

Patients with obstructive airway disease may not allow a complete exhalation so that lung volume comes down to functional residual capacity when intubated, as the loss of elastic tissue allows the bronchioles to collapse before some of the respiratory units have emptied to the size allowed by the applied PEEP. This results in the gradual development of additional PEEP at the alveolar level that is not reflected at the pressure sensor of the ventilator during a normal respiratory cycle; this has been variously termed “dynamic hyperinflation” or “auto-PEEP.” This problem also reduces preload, and may thereby produce hypotension or arterial hypoxia. Paradoxically, it is treated by increasing the level of PEEP applied at the ventilator. This maneuver keeps the bronchioles open longer during expiration, decreasing auto-PEEP and thereby improving preload.

TABLE 4.3. Measures for the prevention of ventilator-associated pneumonia

Item	Comment
Head and upper body positioning to decrease risk of aspiration	Greater than 30 and preferably 45 degrees; must be balanced against concerns for cerebral perfusion
Endotracheal tube	Oral rather than nasal to decrease the risk of sinusitis; consider using a tube designed for continuous aspiration of subglottic secretions if prolonged intubation is anticipated (6)
Gastric tubes	Oral rather than nasal to decrease the risk of sinusitis
Nutrition	Institute enteral nutrition as soon as possible
Choice of gastrointestinal bleeding prophylaxis	Although sucralfate may have a slightly lower rate of associated ventilator-associated pneumonia, it is not as effective as H₂ blockers
Selective gastric decontamination	Still debated; potential improvements in pneumonia rate may be counterbalanced by the emergence of resistant microorganisms

Pneumonia

Patients with pneumonia come to the attention of the neurointensivist in one of three ways. The most common is the development of ventilator-associated pneumonia (VAP) (or other nosocomial pneumonia) in patients with any critical illnesses of the nervous system. The second is the occurrence of aspiration in patients whose acute neurological disease impairs airway control; this may also arise at the time of an emergent intubation. The third is the patient with a chronic neurological disease, such as motor neuron disease, who develops pneumonia as a reflection of progressive loss of airway control.

Prophylaxis

The intensive care unit routine should include an established protocol for the prevention of VAP. Extensive research in the past two decades confirms that most instances, as with other forms of pneumonia, result from aspiration. The measures summarized in Table 4.3 should be employed to the extent tolerated by the patient. It should be commented that this problem is also probably ubiquitous among stroke patients and some form of assessment of the safety of swallowing is highly recommended for these individuals before feeding occurs.

Diagnosis

Although the diagnosis of VAP seems straightforward, studies of this condition have been difficult to perform because of disagreements about diagnostic criteria and appropriate microbiologic techniques. This diagnosis is suspected because of the presence of fever and purulent sputum. However, these findings are frequent in patients with artificial airways, and often represent tracheobronchitis rather than pneumonia. The lack of evidence that antibiotic treatment affects the natural history of patients with tracheobronchitis, coupled with the need to reduce unnecessary antibiotic usage to reduce both costs and the selective pressure driving antibiotic resistance, suggests that antibiotics not be instituted for this problem alone.

Conversely, recent studies establish that early and microbiologically correct treatment of pneumonia improves outcome (7). Thus, the strategy that appears to be optimal is one that correctly identifies patients with pneumonia, covers the important microorganisms for that patient group, and allows early tailoring or termination of therapy as appropriate.

Most neurocritical care patients are too ill for routine chest films. Portable films usually suffice, but evanescent infiltrates often are noted. If these infiltrates are coupled with fever and purulent sputum, clinicians often begin presumptive treatment for pneumonia, which continues for days or weeks, even when the infiltrates are absent on the next film. Therefore, a strategy that terminates antibiotic therapy if the chest radiograph clears rapidly is most appropriate. Although chest computed tomography (CT) scanning is more precise than plain radiography for the diagnosis of pneumonia, its utility for this purpose remains to be determined.

A protocol such as that presented in Table 4.4 is useful to address all of these concerns.

Pulmonary Embolism

Patients with critical illness of the nervous system seem particularly predisposed to develop deep venous thrombosis (DVT) and pulmonary embolism. Most such patients are immobilized by a neurological deficit, coma, hemiparesis, or spinal instability. Some have underlying disorders, which include hypercoagulability; others have contraindications to anticoagulation for the prophylaxis of pulmonary embolism. A few have cardiac or pulmonary problems predisposing them to rightto-left shunting. Venous thrombosis may lead to systemic as well as pulmonary embolism in this group.

The optimal prophylactic regimen in these patients is under constant discussion. For patients without a contraindication to anticoagulation, a combination of pharmacologic and mechanical therapies is probably optimal. Although the use of low molecular weight heparins for this purpose has steadily increased since their approval, enthusiasm for their use must be tempered by the recognition that the lack of a readily available technique for measurement of their effects may result in excessive anticoagulation in some patients. Many neurointensivists favor weight-adjusted-dose subcutaneous unfractionated heparin, or simply give unfractionated heparin subcutaneously every 8 to 12 hours, depending on the patient’s habitus. The multiplicity of opinions regarding anticoagulation in various clinical settings is beyond the scope of this text; therefore, the reader should consult recent guideline statements such as that of the American College of Chest Physicians (available at http://www.chestnet.org/health.science.policy/quick.reference.guides/antithrombotic/6th_antithrom_qrg.pdf).

The diagnosis of uncomplicated DVT requires a high index of suspicion. Physical findings such as lower extremity edema, a palpable venous cord, or Homan’s sign may suggest the diagnosis but are frequently absent. Upper extremity venous thromboses are increasingly recognized, especially in patients with central venous catheters or immobilized or paretic extremities. If the involved vein has been used for venous access, one should consider the possibility of septic thrombophlebitis. Otherwise, systemic signs of inflammation
usually are lacking. Although DVT is sometimes detected in the course of a search for the etiology of a fever, neither venous thrombosis nor pulmonary embolism are likely to cause substantial fever; hence, the search for a source of fever should not end with the discovery of a clot (10).

TABLE 4.4. Protocol for the diagnosis and management of ventilator-associated pneumonia (VAP)

				0 Points	1 Point	2 Points
1.	Suspect VAP in patients mechanically ventilated for ≥48 hours if
	a.	the chest radiograph or chest CT shows a new or progressive and persistent infiltrate AND
	b.	at least one of the following is present:
		1.	Fever (temperature >38°C)
		2.	Leukopenia (<4,000 white blood cell [WBC]/cm3) or leukocytosis (≥12,000 WBC/cm3)
		3.	Altered mental status in patients ≥70 years AND
	c.	at least one of the following:
		1.	New onset of purulent sputum
		2.	New onset or worsening of cough, dyspnea, or tachypnea
		3.	Rales or bronchial breath sounds
		4.	Worsening gas exchange
2.	In patients suspected of having VAP by these criteria, calculate the clinical pulmonary infection score for day 1 (CPIS-1) (8):
Temperature (°C)				36.5-38.4	38.5-38.9	≤36 or ≥39
WBC count				4-11 K	<4 K or >11 K plus one additional point for >50% bands
Secretions				None	Nonpurulent	Purulent
PaO₂/FiO₂ ratio				>240 (no points if PaO₂/FiO₂ ratio abnormal for another reason, such as the acute respiratory distress syndrome)		<240 and no other etiology
Radiographic abnormalities				None	Diffuse or patchy	Localized
3.	Obtain blood cultures and sputum for quantitative culture (QC) according to the preferred local technique (e.g., endotracheal aspirate mixed with diluted acetylcysteine, or protected brush bronchoscopy, or bronchoalveolar lavage).
4.	Initiate appropriate antibiotic therapy for 3 days, based on the organisms and susceptibilities of the intensive care unit. Because initial adequate antibiotic coverage is an important determinant of outcome, this choice should be adequately broad (e.g., consider two agents active against Pseudomonas aeruginosa) at this point, and narrowed on day 3.
5.	On day 3, calculate CPIS-3:
				0 Points	1 Point	2 Points
Temperature (°C)				36.5-38.4	38.5-38.9	≤36 or ≥39
WBC count				4-11 K	<4 K or >11 K plus one additional point for >50% bands
Secretions				None	Nonpurulent	Purulent
PaO₂/FiO₂ ratio				>240 (no points if PaO₂/FiO₂ ratio abnormal for another reason, such as the acute respiratory distress syndrome)	<240 and no other etiology
Radiographic abnormalities				None	Diffuse or patchy	Localized
Progression of infiltrate				No		Yes (and no other etiology)
	a.	If QC yields <10⁵ cfu/mL and CPIS-3 ≤ 4, discontinue antibiotics (9).
	b.	If QC yields >10⁵ cfu/mL and/or CPIS-3 > 4, continue antibiotics; adjust choice of antibiotics according to culture and sensitivity results.
6.	On day 6, assess need for prolonged antibiotic therapy. Unless indicated by a slow clinical response, discontinue antibiotics on day 7.
Protocol developed by the ventilator-associated pneumonia medical management team of the University of Virginia (courtesy of Thomas P. Bleck, M.D.).

The commonly employed tests for lower extremity DVT include ultrasound examination and impedance plethysmography. In the upper extremity, the evaluation often is limited to ultrasound. These techniques are reliable for the diagnosis of thrombosis in the extremities and the central veins of the chest; however, useful techniques for the diagnosis of pelvic thrombi are lacking. We have taken the position that screening for venous thrombosis should be undertaken before the application of “air boots” unless the patient has been ambulatory in the day or two before the neurological disability.

Measurement of the serum concentration of fibrin degradation products by newer sensitive techniques is a useful adjunct to the diagnosis of DVT or pulmonary embolism. However, patients with trauma or recent surgery have elevated levels of these products, which may then be difficult to distinguish from thrombosis. Thus, for many patients in neurocritical care units, this test has good specificity but poor predictive value (11).

The suspicion of pulmonary embolism in neurocritical care practice is typically raised by the abrupt development of tachypnea and difficulty with oxygenation. The main alternative consideration is airway plugging. Classic findings for embolism such as hemoptysis, chest pain, or a pleural friction rub are rarely present. The commonest electrocardiographic change owing to pulmonary embolism is sinus tachycardia; however, more suggestive findings include acute right axis deviation, a new right bundle branch block, or the development of an S1Q3T3 pattern. With emboli of sufficient size and number, the pulmonary artery (PA) pressure becomes elevated; this can be detected as a change in the pressure recorded from a PA catheter, or from an echocardiographic estimate of PA pressure. The diagnosis is typically confirmed with a radiographic study. Ventilation-perfusion scanning was the traditional test for this purpose, but is being rapidly supplanted by spiral CT scanning. Pulmonary angiography is infrequently necessary for the detection of large thrombi, but may still be required to diagnose smaller clots and it offers a potential avenue for thrombolytic treatment if the embolism is found to be very sizable.

Therapy for DVT and pulmonary embolism is often limited in neurocritical care practice by the risks of systemic anticoagulation. If no contraindications exist, one may choose either unfractionated heparin by infusion or intermittent injections of low molecular weight heparin. Although the latter option has become more common, the risks of excessive anticoagulation in patients with acute nervous system disease, the attendant difficulty of measuring the effect of low molecular weight heparin and the lack of an agent to reverse its effects often argue in favor of unfractionated heparin therapy that we still use. Weightbased nomograms for heparin dosing are available; in general, however, one should initially begin with a heparin dose that is slightly below that predicted. If the patient’s risk of hemorrhage is low, one may choose to administer a 5,000 U initial dose of heparin. From our own experience we emphasize that it is imperative to stop the heparin infusion if a PTT above the measurable limit found (100 or 120 s in most laboratories), and to resume at a lower rate when the test value has returned to a measurable range. Treatment with enoxaparin or other low molecular weight heparins is easier to manage, but its effects are more difficult to measure and it is more difficult to reverse if hemorrhagic complications ensue.

Therapeutic anticoagulation carries substantial bleeding risks in patients who have recently suffered CNS hemorrhages or have had recent intracranial or spinal trauma or surgical procedures. Data to quantitate these risks are lacking; however, many avoid systemic anticoagulation for the first 3 days after an ictus. Should patients develop DVT or pulmonary embolism during this period, or have another
contraindication to anticoagulation (e.g., active peptic ulcer disease), the best option may be placement of an inferior vena cava filter. Such patients should still be anticoagulated when possible (12).

The duration of anticoagulation depends on many factors, but typically is continued for 6 months. Warfarin therapy is substituted after the patient is on a therapeutic heparin regimen and no surgical procedures (e.g., tracheostomy) are anticipated. When warfarin is instituted, the older practice of giving large doses of warfarin for the first few days of therapy should be discouraged, because the measurement of prothrombin time is highly dependent on factor VII levels. The international normalized ratio thus calculated may suggest that the patient has been adequately anticoagulated when only the shortest halflife vitamin K-dependent factors actually have been reduced. Furthermore, concentrations of the anticoagulant proteins C and S fall before that of factor II, which may leave the patient in a relatively procoagulant state at the time heparin is discontinued.

Patients with pulmonary emboli hemodynamically significant enough to cause hypotension may require volume resuscitation and vasopressor therapy. Those not rapidly responsive to these measures should be considered for either thrombolytic therapy or thrombectomy; however, most patients in neurocritical care units have relative contraindications to both of these procedures. If such a patient is dying from hemodynamically significant pulmonary embolism, one may still wish to consider these options.

Severe Atelectasis and Mucus Airway Obstruction

Some degree of atelectasis is ubiquitous among patients receiving mechanical ventilation, but large areas of atelectasis, or collapse of segments or entire lobes, often produces substantial difficulties with oxygenation. This often results from mucous plugging of the airways, a problem to which neurocritical care patients seem commonly predisposed. Perhaps this reflects autonomic dysfunction, although the precise etiology is unknown. Although the pulmonary circulation normally responds to this problem by decreasing local perfusion (hypoxic vasoconstriction), many patients with critical diseases of the nervous system seem particularly predisposed to this problem, resulting in more severe hypoxemia than one expects.

Attempts to treat atelectasis in patients receiving mechanical ventilation often include increasing the FiO₂, as well as enhancing the tidal volume and the PEEP in an attempt to ventilate the collapsed regions. Delivering a higher concentration of oxygen is useful as long as the other portions of the lung are normal, and the extent of pulmonary arteriovenous shunting is not too great. Indeed, one can assess the relative proportion of shunting versus ventilation-perfusion mismatching by observing the degree of correction of hypoxemia with 100% inspired oxygen, only the latter being corrected by this maneuver. Some shunting is to be expected with airway plugging and atelectasis, but early on the majority of the difficulty relates to mismatching. Increasing the inspired oxygen fraction should be instituted immediately in order to avoid systemic hypoxemia while measures are instituted to correct the underlying problem. However, increasing the tidal volume only overdistends the other areas of the lung, and by itself does not improve oxygenation. Increasing the PEEP may recruit areas of microatelectasis and usually thereby improve oxygenation, but will not reinflate larger regions of collapse that are lost to reexpansion once their surfactant coating has been reabsorbed.

Physical measures such as suctioning, postural drainage, and external percussion sometimes are effective in relieving the obstruction producing atelectasis, but some patients require bronchoscopic suctioning and lavage. Patients with reactive airway disease may benefit from inhaled β-agonist treatments to relieve obstruction in smaller airways. Occasional patients with substantial secretion volumes also may benefit from either inhaled or systemic anticholinergic agents, but one must
be wary of producing secretions that become too tenacious to clear. Once the obstruction is relieved, maintaining the patient on higher PEEP levels to increase functional residual capacity may help to prevent this problem. Although some have suggested that ventilation with very high tidal volumes (e.g., 20 mL/kg) may help to prevent this condition (and speed weaning) in patients with normal lungs, this should be considered unproved at present. We prefer to ventilate most of these patients with sufficient pressure support to achieve tidal volumes of 7 to 8 mL/kg if they are able to trigger the ventilator, or a similar tidal volume in a controlled volume mode if they are unable to do so.

Neurogenic Ventilatory Failure

Patients with neurogenic ventilatory failure comprise one of the most important and interesting groups in neurocritical care practice. The major causes of this problem are summarized in Table 4.5 and a more detailed account of diaphragmatic failure owing to neuromuscular conditions is discussed in more detail elsewhere in this volume, particularly the one on Guillain-Barré syndrome (Chapter 18). The conditions included in the table do not represent a complete list, but are intended to provide examples of each of the types of neurogenic ventilatory failure (13).

Regardless of the etiology, patients respond to reduced ventilation in one of two characteristic ways, depending on the ability of the brainstem to sense and respond to pH, and the ability of the effector mechanisms to operate, extending from the medulla to the cervical spinal motor neurons, motor nerves, and diaphragmatic neuromuscular junction. It should be noted that extracellular pH drives this response; the PaCO₂ is not an independent driver but rather exerts its influence through the pH. Local (brainstem) or systemic acidosis should trigger an increase in minute ventilation, whereas alkalosis does the opposite. If the effector mechanisms fail completely, the patient becomes apneic. The remainder of this discussion assumes that brainstem commands are able to affect the respiratory musculature.

If the brainstem is unable to sense or respond to a fall in pH, the patient’s minute ventilation is not increased appropriately, and CO₂ retention occurs; however, the actual PaCO₂ must be interpreted in terms of the patient’s overall acid-base balance. If the brainstem is functioning normally but the effectors are impaired, the patient’s response depends in part on the rate of progression of weakness. Abrupt failure of the peripheral mechanisms, as in neuromuscular junction blockade, produces a progressive rise in PaCO₂ (and a concomitant fall in pH). Rapidly progressive disorders such as Guillain-Barré polyneuropathy or myasthenic crisis produce weakness that increases over hours or days. In this setting, the initial fall in tidal volume initially elicits a compensatory increase in that efferent signal in an attempt to restore the tidal volume to normal. However, as the weakness progresses, inadequate lung expansion causes a degree of atelectasis, which by itself usually is not sufficient to cause hypoxemia. This atelectasis appears to be sensed by afferents from the lung as inadequate stretch, however, which is translated into a signal requesting an increase in minute ventilation. At about the same time, the ventilatory system loses the ability to increase the tidal volume further, and as a consequence all requests for increasing minute ventilation can only be met by increasing respiratory rate. The exact points during the progressive fall in muscle strength that the relative contributions of atelectasis (producing the inadequate stretch signal) and the need for a higher respiratory rate (to insure CO₂ excretion) are exerting greater influence over the brainstem control mechanisms are unclear, and probably vary among patients. However, the consequence is an increase in respiratory rate that typically overshoots the rate required to compensate for the fall in tidal volume; this increases the minute ventilation above that needed for the amount of CO₂ produced and thus results in a slight respiratory alkalosis. The clinician caring for such a patient must not let the finding of a PaCO₂ under 40 mm Hg
produce a false sense of security. One must interpret the PaCO₂ in the context of the respiratory rate and the vital capacity, also recognizing that weakness of the cranial musculature may impair the patient’s ability to form a tight seal around the spirometer mouthpiece and hence make the vital capacity artifactually low.

TABLE 4.5. Major causes of neurogenic ventilatory failure

Type	Anatomic class	Representative entities	Clinical manifestations
Type IIa	Central failure to respond to falling pH (inadequate respiratory rate and inadequate tidal volume)	Medullary infarction or other lesion disrupting the nucleus of the solitary tract	Apnea, if complete; CO₂ retention if incomplete
		Narcotic overdose or other intoxication	Absent or inadequate response to falling pH and PaO₂
Type IIb	Disconnection of medullary efferents from lower motor neurons	Cervical spine injury or other lesion between lower medulla and C4 producing diaphragmatic paralysis	Apnea, if pathways are completely disrupted; tachypnea with inadequate tidal volume and consequent rise inPaCO₂ if incomplete; attempted use of accessory muscles
		Cervical spine injury or other lesion between C4 and T6 producing paralysis of parasternal intercostal muscles	Paradoxical respiration; use of accessory muscles; rise in PaCO₂
Type IIc	Disorders of lower motor neurons in spinal cord	Amyotrophic lateral sclerosis and other motor neuron diseases	Slowly progressive weakness resulting in CO₂ retention and compensatory metabolic alkalosis
		Tetanus	Acute or subacute spasticity of musculature causing acute respiratory acidosis; superimposed airway obstruction
Type IId	Disorders of peripheral nerve	Acute inflammatory polyradiculoneuropathy (Guillain-Barré syndrome)	Progressive fall in tidal volume with compensatory rise in respiratory rate (often overcompensate early in the disease)
		Chronic inflammatory polyneuropathy	Ventilatory failure is uncommon
		Diphtheric neuropathy
		Porphyric neuropathy
		Tick paralysis	Dermacentor tick produces neuropathy
		Ciguatera
		Saxitoxin (and other nerve toxins)
		Thallium intoxication
		Acute hyperkalemic paralysis
		Heavy metals producing peripheral neuropathy	Arsenic
		Critical illness polyneuropathy
		Motor neuropathy of critical illness
Type IIe	Disorders of the neuromuscular junction	Myasthenia gravis	Similar respiratory problems to acute inflammatory polyradiculoneuropathy
		Botulism	Associated autonomic disturbances are often prominent from the outset
		Organophosphate intoxication	Carbamate insecticides and nerve agents
		Hypermagnesemia
		Tick paralysis	Ixodes tick toxin produces neuromuscular junction dysfunction
		Snake bites
		Prolonged effect of drugs producing therapeutic neuromuscular junction blockade	Pancuronium
Type IIf	Disorders of respiratory muscles	Rhabdomyolysis Polymyositis/dermatomyositis
		Acid maltase deficiency
		Carnitine palmityl-transferase deficiency
		Necrotizing myopathy of critical illness
		Thick filament myopathy
		Depolarization failure
		Nemaline rod myopathy
		Acute hypokalemic paralysis
		Acute hypophosphatemic paralysis
		Stonefish myotoxin poisoning
		Barium poisoning
		Trichinosis

Eventually, progressive weakness results in a further decline in the tidal volume, for which the patient is unable to compensate with further elevations in respiratory rate; as the minute ventilation falls, the PaCO₂ increases, with a concomitant fall in pH. Such patients usually require mechanical ventilation, although one can opt to permit hypercapnia in the patient who maintains good airway control and intact consciousness. On rare occasions, we have allowed such patients to reach PaCO₂ values in the 90- to 100-mm Hg range while waiting for the effect of therapy (e.g., i.v. immunoglobulin therapy for myasthenic crisis) (14). However, such patients must be managed in an intensive care unit where any decompensation will be discovered and managed immediately.

Weakness of the cranial musculature is a frequent, although not invariable, accompaniment of this decline in ventilatory strength. As a consequence, endotracheal intubation may be required for airway control (to prevent aspiration and facilitate suctioning) before the patient requires mechanical ventilation. When the time comes to intubate the patient, recall that the extra work of breathing imposed by the endotracheal tube often forces the patient to depend on mechanical ventilation as well. Because the resistance imposed by the tube is directly proportional to its length but proportional to the inverse fourth power of its radius, one should place the largest diameter tube that comfortably fits the patient’s larynx. If neuromuscular junction blockade is required to intubate a patient with neuromuscular respiratory failure, the action of the agent used likely lasts much longer than the duration to which one is accustomed. Although the initial sedation and neuromuscular junction blockade are active, the patient requires mechanical ventilation with a fixed rate. When these medications have worn off, the patient may or may not need mechanical ventilation. Five cm H₂O of CPAP is often useful to prevent atelectasis in this group of patients, but one must observe the patient carefully for signs of fatigue because CPAP slightly increases work of breathing, in addition to the increase imposed by the endotracheal tube.

Patients who require mechanical ventilation because of respiratory muscle weakness but who maintain airway protection may be candidates for noninvasive mechanical ventilation. In general, patients with rapidly progressive neurogenic respiratory failure (e.g., acute inflammatory polyradiculoneuropathy) do not tolerate noninvasive ventilation well and are usually better candidates for endotracheal intubation. Conversely, patients with slowly progressive neuromuscular diseases who need short-term mechanical ventilation (e.g., during treatment of an intercurrent illness) occasionally do very well with noninvasive ventilation by mask. Such patients must always have someone in attendance in case their tightly fitting mask should become filled with vomitus. Although some patients also may be managed with negative pressure ventilators (e.g., cuirass ventilators), these devices are more commonly used for long-term home ventilation.

Once the patient with neurogenic ventilatory failure is committed to positive pressure mechanical ventilation via an endotracheal tube, five basic principles must be kept in mind. First, if the patient has developed ventilatory failure over several days or weeks, renal compensation for the respiratory acidosis has produced a high serum bicarbonate concentration. One should not attempt to correct the elevated PaCO₂ immediately, because this produces a severe systemic alkalosis. Rather, one should achieve and maintain a minute ventilation that results in a pH near 7.50; this triggers renal loss of bicarbonate and other bases, eventually bringing the serum bicarbonate back to the normal range. Because the minute ventilation necessary to do this is often quite low, one usually selects a reasonable
tidal volume (e.g., 6 to 8 mL/kg) at a low rate; lower tidal volumes may promote atelectasis.

Second, there is no clear value to “resting” patients by using full ventilatory support, even at night, with the exception of selected patients with chronic obstructive lung disease. One should make maximal use of the patient’s own respiratory capabilities in order to prevent muscle deconditioning to the extent possible. Pressure support ventilation is usually the most easily tolerated ventilator mode for these patients, because it supplies a high gas flow rate at the start of the breath. The optimal level of pressure support is generally that producing a tidal volume in the 6- to 8-mL/kg range, a respiratory rate below 20/min, and no signs of distress (diaphoresis, tachycardia, nasal flaring, and use of accessory muscles of respiration). However, there is no evidence to prove that one mode of ventilation leads to faster weaning than another; if the patient seems more comfortable when a volume mode of ventilation is employed, one should consider using that mode.

Third, at a time when weaning seems feasible, it is useful to “challenge” the patient once or twice daily to ensure that weaning from mechanical ventilation proceeds as quickly as possible. One can either decrease the level of pressure support, or introduce periods (e.g., 1 hour at a time) of CPAP in place of mechanical ventilatory support. The patient should be prepared for these changes by careful explanation and reassurance. If the patient requires sedation to tolerate the ventilator, this should also be reduced or discontinued daily to insure that changes in the patient’s tolerance or accumulation of medication have not produced too deep a level of sedation.

Only gold members can continue reading. Log In or Register to continue