Chapter 31. The Pathophysiology and Differential Diagnosis of Acute Respiratory Failure

• Type I respiratory failure, characterized by severe, oxygen-refractory hypoxemia, is caused by shunting due to airspace filling.
• When blood transport of oxygen is inadequate, treatment includes optimizing cardiac output, hemoglobin concentration, and arterial saturation, and lowering oxygen consumption.
• Optimizing does not mean maximizing, and the end point of each therapeutic approach needs to be selected for the individual patient.
• Type II respiratory failure is characterized by alveolar hypoventilation and increased PCO2, caused by loss of CNS drive, impaired neuromuscular competence, excessive dead space, or increased mechanical load.
• Type III respiratory failure typically occurs in the perioperative period when factors that reduce functional residual capacity combine with causes of increased closing volume to produce progressive atelectasis.
• Type IV respiratory failure ensues when the circulation fails and resolves when shock is corrected, as long as one of the other types of respiratory failure has not supervened.
• Liberation from mechanical ventilation is enhanced by identifying and correcting the many factors contributing to increased respiratory load and decreased neuromuscular competence.

Respiratory failure (RF) is diagnosed when the patient loses the ability to ventilate adequately or to provide sufficient oxygen to the blood and systemic organs. Urgent resuscitation of the patient requires airway control, ventilator management, and stabilization of the circulation, while effective ongoing care for the patient with RF necessitates a differential diagnosis and therapeutic plan derived from an informed clinical and laboratory examination supplemented by the results of special ICU interventions. Recent advances in ICU management and monitoring technology facilitate early detection of the pathophysiology of vital functions, with the potential for prevention and early titration of therapy for the patient's continual improvement. The purpose of this chapter is to provide an informed, practical approach to integrating established concepts of pathophysiology with conventional clinical skills. This chapter does not provide a course in pulmonary physiology nor a comprehensive review of how to treat respiratory failure. Rather, it attempts to provide a conceptual framework of principles useful in approaching the patient with RF, first by discussing an approach to tissue hypoxia, and then by describing the mechanisms causing four types of RF, showing how correcting each derangement allows the patient to resume spontaneous breathing effected by respiratory muscles that are not fatigued.

Patients with RF are susceptible to anaerobic metabolism, either because they deliver inadequate O2 to their systemic organs or because their tissues develop an abnormal inability to extract oxygen from the blood.1 During air breathing, arterialized blood leaves the normal alveoli with a partial pressure of oxygen (PaO2) of about 100 mm Hg. When the hemoglobin concentration is 15 g%, arterial O2 content (CaO2) is about 20 mL per 100 mL blood on the fully saturated hemoglobin and about 0.3 mL in physical solution. Accordingly, a cardiac output (Q̇t) of 5.0L/min transports approximately 1000 mL/min of O2 to the tissues (transport of oxygen; Q̇O2). There, tissue metabolism (oxygen consumption; V̇O2) extracts 250 mL/min, so 5.0 L/min of mixed venous blood returns to the lungs with 750 mL/min of O2, or a mixed venous O2 content (Cv̅O2) of 15 mL per 100 mL blood. Accordingly, the normal extraction fraction [EF = V̇O2/Q̇O2 = (CaO2 − Cv̅O2)/CaO2] is 0.25. Because this O2 content corresponds to 75% O2 saturation (15/20), mixed venous PO2 (Pv̅O2) is 40 mm Hg, as determined by the oxyhemoglobin dissociation curve for normal venous pH, partial pressure of carbon dioxide (PCO2), and temperature. Figure 31-1 plots the relationships between V̇O2 (left ordinate) and EF (right ordinate) with Q̇O2 (abscissa) using these values normalized to a body size of about 65 kg: V̇O2 is 4 mL/kg per minute and Q̇O2 is 16 mL/kg per minute as plotted on the continuous lines.

In many patients with RF, the O2 transport to the tissues is reduced by abnormally low cardiac output, hemoglobin, or O2 saturation. Consider the effects of acute myocardial injury or hypovolemia that reduces Q̇t to 2.5 L/min and Q̇O2 to 8 mL/kg per minute. To maintain the V̇O2 necessary for aerobic metabolism, the tissues must extract 250 mL/min from half the blood flow, so Cv̅O2 decreases to 10 mL O2 per 100 mL blood and EF increases to 0.50. Because this value corresponds to 50% saturation (10/20) of the normal hemoglobin concentration, Pv̅O2 is 27 mm Hg. When cardiac output is returned toward normal with vasoactive drug or volume therapy, Pv̅O2 rises again. In another patient with normal cardiac output (5 L/min) but severe arterial hypoxemia (PaO2 = 40 mm Hg, O2 saturation = 75%, CaO2 = 15 mL O2/100 mL blood), CvO2 must decrease to 10 mL per 100 mL blood to provide the tissues with 250 mL/min of O2; again, Pv̅O2 decreases to 27 mm Hg. When cardiac output increases in response to hypoxia, Pv̅O2 increases again. In a third patient with normal Q̇t and PaO2 but with reduced concentration of hemoglobin (7.5 g/100 mL blood), CaO2 is reduced to 10 mL per 100 mL blood. Accordingly, Cv̅O2 and Pv̅O2 must decrease to 5 mL per 100 mL blood and 27 mm Hg, respectively, to maintain aerobic metabolism, and these venous values increase again with greater cardiac output or hemoglobin concentration. In each case, V̇O2 was maintained as Q̇O2 decreased, creating the horizontal continuous line in Fig. 31-1 indicating that VO2 is independent of O2 delivery in this range; V̇O2 was constant because EF increased in the hyperbolic manner indicated by the continuous line relating EF to Q̇O2.

These considerations illustrate that one end point of reduced O2 transport in the blood is reduced Pv̅O2. Since Pv̅O2 approximates the PO2 adjacent to the exchange vessels in the tissues, it is the driving pressure for O2 diffusion from the capillaries to the metabolizing cells. When Pv̅O2 falls too low, insufficient O2 diffuses to maintain aerobic metabolism, and the cells begin to produce lactic acid as the end point of anaerobic metabolism.2 This is illustrated in Fig. 31-1 by the decrease in V̇O2 as Q̇O2 is reduced below 6 mL/kg per minute; this supply dependency of VO2 occurs when EF exceeds 0.67 and tissue O2 extraction can no longer increase along the hyperbolic continuous line relating EF to Q̇O2. Accordingly, reduced Pv̅O2 and increased serum lactate (or falling pH with unchanged partial pressure of arterial carbon dioxide [PaCO2]) are indications of tissue hypoxia. This improves with therapy increasing CaO2 (by increasing hemoglobin or O2 saturation) and increasing cardiac output. In many critically ill patients, two or three of these factors reducing O2 transport to the tissues coexist, so attention to optimizing all these (Q̇t, hemoglobin concentration, and O2 saturation) is reasonable in the hypoxic patient. This description of the effects of diminished Q̇O2 on Pv̅O2 and hence the utility of Pv̅O2 as a monitor of the adequacy of Q̇O2 helps one understand the benefits of early goal-directed therapy guided by this measurement in the treatment of patients with shock3 (see Chap. 21).

Optimizing QO2

Of course, optimizing does not mean maximizing, and the end point of each therapeutic approach needs to be selected for the individual patient. Patients with low cardiac output due to heart disease may not tolerate infusions of packed erythrocytes even though their tissue hypoxia is made worse by concurrent anemia. Yet thoughtful integration of packed cells within the therapy of plasma volume reduction is often helpful in such patients and may prevent anaerobic metabolism at a time when cardiac output cannot be increased to adequate levels. In other patients with severe arterial hypoxemia and O2 desaturation due to acute hypoxemic respiratory failure, tissue hypoxia may be relieved by increasing cardiac output and hematocrit. Yet this is often associated with higher central blood volume and pulmonary vascular pressures that increase pulmonary vascular leakage unless vasoactive drugs, diuretics, and fluid restriction are used concurrently. Again, thoughtful integration of the three approaches to therapy of tissue hypoxia provides the optimal level of circulating hemoglobin and cardiac output while reducing rather than aggravating the pulmonary edema. Some patients with chronic severe anemia (e.g., chronic renal failure) become acutely ill with low cardiac output and/or hypoxemic respiratory failure. Their tissue hypoxia is often ameliorated by prompt, transient increases in their hemoglobin concentration without circulatory overload, as by plasmapheresis. Yet the institution of this therapy, like the others mentioned above, has complications that must be weighed against the likely benefit in that patient at that time. Accordingly, this approach to therapy of reduced blood O2 transport implements early each of the three major interventions (Q̇t, hematocrit, and arterial oxygen saturation [SaO2]) in a combination best suited to the condition of each patient.2,3

Dissolved O2 contributes a very small amount to Q̇O2. Yet, in critical hypoxemia, raising the fraction of inspired oxygen (FiO2) to maximal values may be effective.2 Consider again the patient with acute myocardial infarction without lung disease in whom low cardiac output has lowered Pv̅O2 to 27 mm Hg during air breathing. Even though the hemoglobin is fully saturated, CaO2 may be increased by 1.7 mL per 100 mL blood when PaO2 is increased to 650 mm Hg by ventilation with O2. Then Cv̅O2 increases from 10 to 11.7 mL O2 per 100 mL blood, raising mixed venous saturations to 58% and Pv̅O2 to about 34 mm Hg; of course, if anaerobic metabolism existed before oxygen therapy, Cv̅O2 may not increase as much as CaO2 because V̇O2 increases with oxygen. These changes tend to diminish tissue hypoxia and the adverse consequences of anaerobic metabolism by an amount equivalent to that achieved by a 1 g% increase in hemoglobin or a 0.2 L/min increase in cardiac output and so complement a combined approach to hypoxia.2 Increasing FiO2 may be effected by nasal prongs to deliver O2 at 1 to 5 L/min (FiO2 .21 to 0.4), by rebreathing masks (FiO2 .21 to 0.6), or by head tent (FiO2 .21 to 0.8). The ranges of FiO2 are to indicate that all methods frequently give no O2 enrichment due to inadequate delivery to the patient (lower limit), and that the amount of O2 delivered is often less than expected (upper limit), even when the O2 delivery system is working properly (Table 31-1).

Reducing VO2

With its attendant risks, tracheal intubation ensures delivery of the highest possible FiO2 and allows another approach to therapy of tissue hypoxia, namely to reduce V̇O2 (Table 31-2). Normally, the work of breathing is very low, but in patients with acute hypoxemic respiratory failure and its associated tachypnea and lung stiffness, V̇O2 of the respiratory muscles alone can approach 100 mL/min.4 Increased work of breathing may result in high V̇O2 in other patients with other causes of restriction such as morbid obesity5 (see Chap. 109). This is illustrated by the interrupted lines in Fig. 31-1 showing that supply dependence of V̇O2 occurs at a high level of Q̇O2 when V̇O2 is tripled (12 mL/kg per minute) by common concomitants of critical illness, even when the tissue ability to extract O2 is normal. Normally, cardiac output increases with V̇O2, as in exercise, to keep Cv̅O2 and Pv̅O2 close to their resting values. Yet consider the effects of such work of breathing in the common circumstance of cardiogenic pulmonary edema when cardiac output may not increase much above 5.0 L/min. Then Cv̅O2 must fall toward 5 mL O2 per 100 mL of blood to deliver the total V̇O2 of 750 mL/min so that Pv̅O2 approaches 22 mm Hg, a level associated with tissue hypoxia and anaerobic metabolism. Relaxation of the respiratory muscles and positive pressure ventilation reduce V̇O2 to 250 mL/min and raise Pv̅O2 to normal with no change in cardiac output.4

Another effect on tissue hypoxia in patients already ventilated is to reduce V̇O2 by cooling the febrile hypoxic patient.6 Consider the patient with pneumonia causing PaO2 of 40 mm Hg (CaO2 = 15 mL%). Then reduction of V̇O2 from 500 to 250 mL/min by sedation, muscle relaxation, and cooling from 40° to 37°C raises Cv̅O2 from 5 to 10 mL O2 per 100 mL of blood. This increase in mixed venous saturation from 25% to 50% would increase Pv̅O2 from 22 to 27 mm Hg in normothermic blood. The left shift of the oxyhemoglobin dissociation curve between 40° and 37°C does not limit oxygen extraction in canine studies of the limits of aerobic metabolism,7 so cooling the febrile patient may be enough to relieve tissue hypoxia in critical situations.

Detecting Anaerobic Metabolism in Respiratory Failure

To illustrate how several clinical interventions have a beneficial effect on tissue hypoxia, this discussion emphasized how Pv̅O2 tracks the changes in Q̇O2. Yet the value of Pv̅O2 at the onset of anaerobic metabolism might vary widely as a result of the Q̇O2/V̇O2 variance among peripheral tissues.8,9 This is especially true in the septic patient when very high Q̇t and Q̇O2 are associated with very high values of Pv̅O2 and lactic acidosis. To the extent that such lactic acidosis arises from anaerobic metabolism, the rise in Pv̅O2 with increased Q̇O2 in the septic patient confounds the utility of Pv̅O2 as a marker of tissue hypoxia. Furthermore, lactic acidosis at high levels of oxygen transport does not necessarily signal pathologic supply dependence of oxygen utilization; rather, the high O2 demands of critical illness may exceed even normal extraction limits from the apparently high but insufficient O2 transport.1 An important observation complicating this problem is that lactic acidosis when Q̇O2 is high does not necessarily indicate anaerobic metabolism, but instead accelerated aerobic glycolysis associated with sepsis-related disturbance of important glycolytic enzymes.10,11 Clinical and experimental studies demonstrate that progressive reduction in Q̇O2 due to hypovolemic or cardiogenic shock is associated with lactic acidemia having a high lactate-to-pyruvate ratio (L/P); yet, in septic shock, the frequently observed lactic acidemia, even at high levels of Q̇O2, is not associated with an increased L/P ratio, for the pyruvate levels have risen in proportion to the lactate levels.12 These observations raise the possibility that metabolic utilization of tissue protein stores in septic shock produces abundant pyruvate in excess of that required for generation of high-energy adenosine triphosphate (ATP) bonds through aerobic glycolysis; then excess pyruvate circulates in normal equilibrium with excess lactate, and clinicians mistake this lactic acidosis for anaerobic metabolism. To the extent that pathologic supply dependence does not occur in patients,13,14 and that the lactic acidosis of sepsis is not anaerobic, the critical care practice of maximizing cardiac output and Q̇O2 confers no benefit on oxygen utilization. Even in the apparent absence of sepsis, the value of Pv̅O2 at the onset of anaerobic metabolism varies widely, and some organs may be anaerobic when Pv̅O2 and Q̇O2 are not worrisome.15 Accordingly, measuring and following changes in Pv̅O2 and venous saturation in conjunction with acid-base status and lactic acid measurements allow deductions concerning the effects of altered Q̇O2 on V̇O2 and aerobic metabolism, but these are nonspecific and subject to errors and uncertainties. It seems reasonable to ensure sufficient Q̇O2 to provide a normal Pv̅O2 in septic patients without maximizing Q̇O2 to treat hypothetical tissue O2 extraction defects,13,14 except where measurement of organ acidosis demonstrates improvement with increased Q̇O2 .15

Hypoxemic versus Ventilatory Failure

A descriptive survey of patients requiring mechanical for respiratory failure reveals four patterns of pathophysiology, each having a predominant mechanism1 (Table 31-3). Intrapulmonary shunt (Q̇s/Q̇t) causes hypoxemia refractory to O2 therapy despite hyperventilation and reduced PaO2 in type I or acute hypoxemic respiratory failure (AHRF).2,16 Primary failure of alveolar ventilation (V̇a) leads to CO2 retention and arterial hypercapnia associated with reduced PaO2; this hypoxemia corrects easily with O2 therapy in type II or ventilatory failure.1,2,17  Figure 31-2 illustrates the different mechanisms for these two common abnormalities in pulmonary O2 exchange.2

Abnormalities in Pulmonary Gas Exchange

It is helpful to recall that the cause of hypocapnia (inadequate V̇a) is often independent of the cause of hypoxemia,1 so the treatment for hypercapnia (raising V̇a) is different from the treatment for hypoxemia (raise FiO2, positive end-expiratory pressure [PEEP]). When CNS depression of the drive to breathe or loss of neuromuscular coupling (see Chap. 66) reduces minute ventilation (V̇e), the CO2 produced at rest each minute (V̇CO2 = 250 mL/min) is added to the reduced V̇a (normally about 4 L/min), raising the alveolar and arterial partial pressures (P) of CO2 (PaCO2 = PaCO2 = k × V̇CO2/V̇a); recall that the P of a gas in the lungs is the product of the gas fraction (Fg) and the dry barometric pressure (Pbar2PH2O), so PaCO2 = (250/4000) × 713 × 0.86, or about 40 mm Hg. Mild alveolar hypoxia develops as the required oxygen uptake (V̇O2) is absorbed from the reduced V̇a (PaO2 = PiO2 – PaCO2/R, where R = V̇CO2/V̇O2), so the consequent arterial hypoxemia is corrected with small increments in inspired oxygen fraction (FiO2). In diseases characterized by airflow obstruction17–19 or lung restriction,20 V̇a is reduced despite normal or increased V̇e because the dead space:tidal volume ratio [Vds/Vt = (PaCO2 − PeCO2)/PaCO2] is increased when large numbers of poorly perfused alveoli are excessively ventilated (high V̇a/Q̇ units). Accordingly, when a patient requires an abnormally large V̇e to maintain a normal PaCO2, the causes are an abnormally increased V̇CO2, Vds/Vt, or both. Hypoxemia develops with airflow obstruction or lung restriction when other alveoli are poorly ventilated in relation to their perfusion (low V̇a/Q̇ units), and the hypoxemia is made worse by low mixed venous oxygen content2,17–20 (see Chap. 39); again, modest increments in FiO2 correct this hypoxemia (see Fig. 31-2, panel A). By contrast, even an FiO2 of 1.0 cannot correct the hypoxemia induced by increased Q̇s/Q̇t (see Fig. 31-2, panel B), and this refractory hypoxemia of AHRF is often associated with increased V̇e and V̇a and so decreased PaCO21 (see Chap. 38).

Abnormalities in Respiratory Mechanics

These two classic types of respiratory failure have distinctly different abnormalities in the mechanics of breathing (Fig. 31-3), while they share mechanisms leading to respiratory muscle dysfunction and fatigue.1 The schematic illustration of the normal respiratory system (see Fig. 31-3A) indicates that spontaneous respiration is effected by the pressure (ΔP) generated by inspiratory muscles to expand the lung and chest wall (ΔV) against their elastance and to cause inspiratory flow (V̇i) past the airways resistance (R). When R is increased in acute-on-chronic respiratory failure (ACRF) or in status asthmaticus,1,22 the ΔP required to breathe often exceeds the strength of the respiratory muscles, resulting in fatigue of the muscles. When such a patient is mechanically ventilated, the peak pressure (Ppeak) generated at the airway opening increases well above the normal value (about 20 cm H2O in Fig. 31-3B, upper left) to 60 cm H2O (Fig. 31-3B, lower right); yet end-inspiratory occlusion allows the airway pressure to return to the normal elastic pressure (Pel = 10 cm H2O in these same panels), confirming that the resistive pressure (Pr = Ppeak − Pel) has increased fivefold from the normal Pr value of 10 cm H2O. By contrast, when airspaces become flooded with edema, pus, or blood in AHRF, the tidal volume is delivered to a smaller number of aerated alveoli, overdistending them to increase Pel and Ppeak.1,23,24 In the lower left panel of Fig. 31-3B, Ppeak is increased to 60 cm H2O because end-inspiratory occlusion pressure has increased to 50 cm H2O; in this instance of increased elastance, Pr (Ppeak − Pel) is normal. This large increase in the ventilator pressure mirrors the increased load on the respiratory muscles during spontaneous breathing; combined with the tachypnea, O2 desaturation, and acidosis in AHRF, this leads to respiratory muscle fatigue.1,21,23,24

Clinical Presentation and Treatment Goals

These mechanical and lung gas-exchange distinctions between AHRF and ventilatory failure are paralleled by substantial differences in the clinical description and presentation of patients with type I and type II respiratory failure (Table 31-3). In the second type, decreased V̇a is caused by reduced drive to breathe (e.g., drug overdoses or head injuries), or by reduced coupling of the adequate or increased drive to breathe to the respiratory muscles (e.g., myasthenia gravis, Guillain-Barré syndrome, amyotrophic lateral sclerosis, botulism, or muscle-relaxing drugs); often the clinical and radiologic examinations of the chest are normal in these patients (see Chap. 66). Alveolar hypoventilation also occurs commonly in respiratory diseases characterized by airflow obstruction and wasted ventilation; PaCO2 increases despite increased CNS drive to breathe, adequate neuromuscular coupling, and increased total ventilation (e.g., status asthmaticus, ACRF, or restrictive pulmonary disease). Wheezing, accessory muscle use, and hyperinflated lungs are common clinical and radiologic features in ACRF and asthma17,18,25,26 (see Chaps. 39 and 40), while crackles and typical interstitial shadows without consolidated airspaces are observed in patients with RF due to pulmonary fibrosis.20 The clinical presentation of each of these categories of hypoventilating patients contrasts markedly with that of patients presenting with type I AHRF, where cyanosis, tachypnea, and refractory hypoxemia lead to early identification of airspace flooding by physical and radiologic examinations3 (see Chap. 38). The differential diagnosis of the airspace flooding leading to Q̇s/Q̇t includes cardiogenic or permeability pulmonary edema,27 pneumonia, and lung hemorrhage,28 each having specific etiologies and therapy.

While specific diagnostic and treatment plans are being implemented, goals of supportive therapy for types I and II RF are quite different. Therapy for patients with AHRF includes four objectives: (1) stabilization of the patient on the ventilator with minimal respiratory work, (2) ventilation with the least tidal volume providing adequate CO2 elimination, (3) addition of the least amount of PEEP effecting 90% saturation of an adequate circulating hemoglobin on a nontoxic FiO2, and (4) cardiovascular management to reduce airspace edema by seeking the lowest pulmonary vascular pressures compatible with an adequate cardiac output and oxygen transport to the peripheral tissues (see Chap. 38 for further discussion). Each of these goals of management of type I AHRF differs from the goals of management of type II ventilatory failure, where the patients with depressed CNS drive or reduced neuromuscular coupling receive adequate ventilation and require minimal oxygen supplementation with careful attention to preventing atelectasis and correcting hypoperfusion until the abnormal neurologic condition resolves; the patients requiring ventilation for airflow obstruction are supported with bronchodilator therapy and ventilator settings that minimize intrinsic PEEP until the airways resistance is reduced sufficiently for the respiratory muscles to achieve adequate ventilation independent from the ventilator (see Chaps. 39 and 40 for further discussion). One reason for identifying the causes of increased respiratory load is to view liberation from mechanical ventilation for patients with type I or type II RF as the mirror image of the cause of RF (i.e., a systematic approach to reverse factors increasing the respiratory load and decreasing the respiratory muscle strength) (Table 31-4).

Perioperative Respiratory Failure

The physician frequently encounters patients in the perioperative period who are unusually susceptible to atelectasis as a primary mechanism causing type III or perioperative respiratory failure.29,30 In general, abnormal abdominal mechanics reduce the end-expired lung volume (↓FRC)31–33 below the increased closing volume (↑CV) in these patients,31,34,35 leading to progressive collapse of dependent lung units (see Table 31-3). The end result can be type I AHRF, or type II ventilatory RF, or both. Yet identification of atelectasis as a distinct mechanism leading to this third type of respiratory failure can be harnessed to prevent lung collapse by reducing the adverse effects of common clinical circumstances promoting reduction in FRC, and of those conditions promoting abnormal airways closure at increased lung volume. Because many of these mechanisms are shared by patients with type I or type II respiratory failure, implementation of approaches to minimize atelectasis should be a part of the management of all patients with respiratory failure.

The principles of preventing or reversing type III perioperative respiratory failure are listed in Table 31-4. Bedside nurses in the ICU turn the patient every 1 to 2 hours from the supine position to the left or right lateral decubitus position; during this time, they provide vigorous chest physiotherapy with pummeling, chest vibration, and endotracheal suction. In patients vulnerable to atelectasis, a fourth position 30° to 45° upright is helpful by reducing the load imposed by the abdomen; also, the addition of sighs, continuous positive airway pressure (CPAP), or PEEP returns the end-expired lung volume to a position above the patient's closing volume.34 Special attention to the treatment of incisional or abdominal pain (e.g., epidural anesthesia or transcutaneous electrical nerve stimulation) and to minimization of the intra-abdominal pressure of ascites or tight bandages helps to prevent atelectasis.32,36 When lobe or lung collapse is detected by physical or radiologic examination, an early, aggressive to re-expansion includes placing the patient in the lateral decubitus position with the collapsed lobe uppermost for vigorous pummeling and suctioning, and then increasing the tidal volume progressively to a pressure limit of 40 cm H2O with end inspiratory pauses. Re-expansion often occurs within 10 minutes and is signaled by a fall in the ΔPel associated with the normal tidal volume at the end of the re-expansion maneuver; if this re-expansion is not confirmed radiologically, repeating these maneuvers after bronchoscopy to clear endobronchial obstructions is indicated. Once re-expansion has occurred, the implementation of increased levels of PEEP and/or sighs often prevents further episodes of atelectasis. Discontinuation of smoking at least 6 weeks prior to elective operations reduces bronchorrhea and atelectasis,37 and avoiding overhydration in perioperative patients especially vulnerable to atelectasis reduces this problem.

Hypoperfusion States Cause Type IV Respiratory Failure

A significant number of ventilated patients fall outside the categories of type I, II, or III respiratory failure. These are the patients who have been intubated and stabilized with ventilatory support during resuscitation from a hypoperfusion state, so type IV RF is most commonly due to cardiogenic, hypovolemic, or septic shock without associated pulmonary problems (see Chaps. 21 and 46). The appropriate rationale for ventilator therapy in these patients who are frequently tachypneic with erratic respiratory patterns is to stabilize gas exchange and minimize the steal of a limited cardiac output by the working respiratory muscles until the mechanism for the hypoperfusion state is identified and corrected.21,38,39 Note that liberation from the ventilator of the patient with type IV RF is simple: when shock is corrected, the patient resumes spontaneous breathing and is extubated. Note further that when patients with type I, II, or III RF suffer a concurrent hypoperfusion state, the causes of reduced blood flow, hypotension, anemia, acidosis, and sepsis need identification and correction as part of the liberation process (see Table 31-4). Accordingly, the physician managing RF frequently employs the principles of diagnosis and management of hypoperfusion states, as discussed in the next section.

The respiratory muscles share with other major muscle groups of the body the characteristic that excessive work leads to fatigue.21,40,41 This concept seems to explain why patients with severe airflow obstruction or airspace flooding ultimately stop breathing, and why patients requiring mechanical ventilation for these and other causes of respiratory failure are unable to breathe independent from the ventilator until the load on their respiratory muscles is reduced, the respiratory muscles become stronger, or both. Note, however, that while this is a useful paradigm with which to manage patients with RF, it is exceedingly difficult to identify fatigue under clinical conditions. Nonetheless, as a rough guide, spontaneous ventilation can be sustained indefinitely when the effort of each spontaneous breath is less than one third the maximal respiratory effort achievable.40,41 In normal patients, the maximum negative inspiratory pressure (MIP) measured at FRC exceeds 100 cm H2O, whereas the work of spontaneous breathing is less than 10 cm H2O, providing considerable respiratory muscle reserve before the conditions of fatigue are approached. In contrast, patients with acute respiratory failure frequently have values of MIP <30 cm H2O, while the load on the respiratory muscles, as measured by the pressure generated by the ventilator during each breath, exceeds 30 cm H2O.22–24 Such values predict that the patient's respiratory muscles will fatigue quickly if spontaneous ventilation were required, a hypothesis easily confirmed in such patients who breathe rapidly and insufficiently when taken off the ventilator.42 Another measure of maximum respiratory effort in the conscious patient is vital capacity (VC). As a rough guideline, when VC is three times the tidal volume (Vt) required to maintain eucapnia and normal pH, respiratory muscle fatigue is unlikely. A corresponding alternate measure of respiratory load is the minute ventilation required to maintain normal PaCO2 and pH. Factors that increase CO2 production, dead space, or metabolic acidosis necessarily increase this ventilation and so promote respiratory muscle fatigue. Such fatigue is often signaled by increased respiratory rate (RR >35 breaths per minute), by paradoxical respiratory motion (the abdomen moves in with inspiration as the fatigued diaphragm is pulled craniad by the negative pleural pressure), and by the patient's unexplained somnolence or decreased responsiveness.21,41 Accordingly, evaluation of the patient's ability to resume spontaneous ventilation includes measurements of MIP, VC, Vt, RR, and V̇e, as well as direct observation of the respiratory motions during a period of spontaneous breathing43,44 (see Chap. 44).

Resting Fatigued Respiratory Muscles

Current evidence and common sense suggest that the treatment for respiratory muscle fatigue is respiratory muscle rest, a strategy that must be balanced in nearly all patients against a thoughtful respiratory exercise program. The timing of the move from respiratory muscle rest to an exercise program is not currently guided by objective criteria identifying fatigue. Accordingly, many physicians confronted with this problem develop empirical guidelines as to the likely presence of respiratory muscle fatigue integrated with the type of early ventilator management necessary for the patient's overall condition. For example, the cardiovascular stability and optimal ventilator management of patients with type I AHRF are frequently enhanced by respiratory muscle rest during the first 6 hours after elective intubation for severe hypoxemia. During this time, the acutely depleted glycogen stores of the resting respiratory muscle are repleted, and the accumulated lactic acid or other metabolites associated with fatigue are washed out;38,39,45 then the patient is ready to move to a respiratory exercise program. By contrast, the patient with ACRF who has developed respiratory muscle fatigue over a longer period of time may require up to 72 hours of respiratory muscle rest before resuming an exercise program free from fatigue. Note that many patients being managed for respiratory muscle rest by mechanical ventilation are actually working as hard or harder than during spontaneous ventilation by breathing actively against the ventilator.46 This is easily detected by clinical examination coupled with observation of the airway pressure, which should rise at the start of each ventilated breath. In many patients, the airway pressure stays at or below zero during inspiration, indicating active inspiration by the patient; the amount and duration of inspiratory effort often can be assessed by the fall in central venous or pulmonary artery pressure with each inspiration. When respiratory rest is indicated, it can be achieved in these circumstances by increasing the inspiratory flow rate, by increasing the minute ventilation, or, if these measures of eliminating the patient's respiratory effort are inadequate, by sedating and paralyzing the patient to ensure respiratory rest.40

Exercising Rested Respiratory Muscles

As soon as the patient with respiratory failure is stabilized on the ventilator, the physician should make a decision whether to rest the fatigued respiratory muscles or to institute a program of respiratory muscle exercise in those patients in whom fatigue is neither evident nor expected. The objective of the respiratory exercise program is to increase the tone, power, and coordination of the respiratory muscles.21,47 The efficacy of each program in increasing tone and power is evaluated by the daily MIP and vital capacity measurements. Coordination is evaluated by bedside observation confirming that the patient's respiratory efforts interact with the ventilator in a manner that is comfortable for the patient. Here the goal is to adjust the ventilator such that the patient receives a breath on demand at a volume and frequency within the ranges expected for that patient off the ventilator. Several different modes of ventilation seek to achieve these goals of increased tone, power, and coordination during the liberation process, and each is described in detail in Chaps. 32, 36, and 44.39

Liberating the Patient from Mechanical Ventilation

This aggressive approach attempts to liberate the patients as soon as possible from mechanical ventilation by using the mode as an exercise program. Then ventilator mode becomes a thoughtful part of a larger program addressing about 50 correctable factors constraining the patient's freedom to breathe (Table 31-4). This effective approach to liberating patients from the ventilator measures and attempts to increase the values of MIP and VC while simultaneously measuring and reducing the respiratory load.47 The left column of Table 31-4 lists correctable factors to reduce the respiratory muscle load. As a general rule, these are the abnormal respiratory mechanics associated with the several types of acute respiratory failure. The right column of Table 31-4 lists correctable factors increasing respiratory muscle strength, first in the context of those many disturbances of the circulation or internal environment most common to patients with type IV respiratory failure. Attempting to liberate patients with hypoperfusion states or hypotension is almost never successful. Correction of these hemodynamic variables complements the correction of anemia, hypoxemia, and acidosis to provide dramatic increases in the objectively measured respiratory muscle strength by MIP and VC. Similarly, attempting to liberate patients who are septic or have body temperatures >38.5°C is often unsuccessful. While these systemic abnormalities are being corrected, it is also helpful to initiate adequate daily protein-calorie nutrition utilizing protein (0.8 g/kg) and nonprotein calories (about 30 kcal/g of protein), of which calories 20% to 50% should be supplied as lipid.48 Elemental malnutrition is corrected by adjustments of serum potassium, calcium, magnesium, and phosphate levels; severe abnormalities of each of these electrolytes is sufficient to cause respiratory muscle fatigue,49,50 so modest abnormalities in the patient already weakened by critical illness may converge to make the patient weaker than necessary. When each of these abnormalities has been corrected, there is some evidence that respiratory muscles may be strengthened by the infusion of aminophylline in doses that achieve serum concentrations associated with effective bronchodilation.51 When all other factors are corrected but the patient remains weak, it is helpful to exclude clinical conditions that occasionally cause reduced respiratory muscle strength. Neuromuscular disease, muscle-relaxing drugs,52 steroids,53 sedatives, coma, and intercurrent cerebrovascular accidents cause obscure reductions in respiratory muscle strength. Often overlooked causes of inadequate respiratory muscle function are subclinical status epilepticus, hypothyroidism, and paralysis of the phrenic nerve on one or both sides after cardiac surgery with cold cardioplegia or other thoracic trauma.47