Management ofthe respiratory system after cardiac surgery depends on multiple factors including preoperative cardiac and pulmonary function, intra- operative ...
- - - - 4~II
• .......---c-rit-i-c8-1C_8_re
_
Respiratory Management after Cardiac Surgery* Michael A. Matthay, M.D., F.C.C.P.;t and Jeanine P. Wiener-Kronish, M.D., F.C.C.P.t
Management ofthe respiratory system after cardiac surgery depends on multiple factors including preoperative cardiac and pulmonary function, intraoperative events, and the patients postoperative hemodynamic status. In recent years, an increasing number of cardiac procedures have been done for patients who are older and have more severe heart disease and more severe chronic obstructive lung disease .1.2 In addition, a significant fraction of these patients have chronic medical problems including systemic hypertension, diabetes mellitus, or chronic renal failure, all ofwhich may complicate postoperative management. This article will first review the effects ofanesthesia, thoracic surgery, and cardiopulmonary bypass on pulmonary function. Then, respiratory management following uncomplicated cardiac surgery will be discussed. The diagnosis and respiratory management of cardiogenic and noncardiogenic pulmonary edema following cardiac surgery will be reviewed. Also, the preoperative and postoperative management of patients with chronic obstructive lung disease will be considered in some detail, as well as the diagnosis and management of unilateral or bilateral diaphragm dysfunction following cardiac surgery. Effects of Anesthesia on Pulmonary Function
It has been well documented that general anesthesia results in approximately a 20 percent decrease in functional residual capacity.3-6 The decrease in functional residual capacity probably is due primarily to the effects of general anesthesia on the shape and motion of the chest wall and diaphragm. 7- 10 Upon induction of general anesthesia, the diaphragm shifts cephalad, especially the dependent portions of the ·From the Departments of ~fedicine and Anesthesia and the Cardiovascular Research Institute, University of California, San Francisco. tSupported in part by a National Institute of Health Pulmonary Vascular SCOR Grant, HL 19155. Reprint requests: Dr. Matthay, 1315-Aloffitt Hospital, University of California Sf; San Francisco 94143-0130
424
diaphragm (Fig 1). This shift may contribute to a decrease in functional residual capacity. The shift in the diaphragm occurs because ofa loss of tonic activity; some work suggests that the anesthesia-induced changes in the chest wall and in diaphragm function may be mediated through effects of anesthesia on the central nervous system. 6 •11 The changes in the chest wall and diaphragm result in alterations in the distribution of the inspired gas without a corresponding adjustment in pulmonary blood flow Thus, low ventilation-perfusion lung units are created which result in a widened alveolar-arterial oxygen difference. There also are some data indicating that atelectasis occurs immediately after induction of anesthesia, and the atelectasis resolves after application of 10 cm H 20 PEE~5.12 This atelectasis may contribute to ~ypox~mia during general anesthesia. It is not clear how long the hypoxemia persists after the termination of general anesthesia. General inhalational anesthetics also inhibit hypoxic The blunting of hypulmonary vasoconstrict~on.4.13 poxic pulmonary vasoconstriction may then contribute to further widening of the alveolar-arterial oxygen difference intraoperatively, particularly in patients who have moderate to severe preoperative ventilationperfusion mismatch or right-to-Ieft intrapulmonary shunts. The deleterious effect of inhalational anesthetics on hypoxic pulmonary constriction causes minimal gas exchange abnormalities in patients with normal lungs. 14 Intravenous narcotics that are used for induction or maintenance of general anesthesia may substantially reduce the postoperative hypoxic and hypercapnic ventilatory drive. Although there is considerable individual variation in the response to general anesthetics and narcotics in terms of the postoperative effect on ventilatory drive, residual effects of narcotics may delay weaning patients from mechanical ventilation. Elderly patients appear particularly susceptible to hypoxemia postoperatively if they have received narcotics. 15 Respiratory Management after Cardiac Surgery (Matthay, Wl6ner-Kronish)
II
I':::::: I.:;:::;::
..............._n_n_n_rl""'"'ln.-..·.~:. . .__ AWAKE SPONTANEOUS
_nn~t
,: ')f ~"
/'-;'
ANESTHETIZED SPONTANEOUS
----...........nnnrr1n PARALYZED
FIGURE 1. Dashed line is the position of the diaphragm at functional residual capaci~ Stippled area represents diaphragmatic excursion during tidal breathing. Note how the diaphragm in the anesthetized subject shifts cephalad, particularly the most dependent parts of the diaphragm (Reprinted with permission from Froese AB, Bryan AC. Effects of anesthesia and paralysis on diaphragmatic mechanisms in man. Anesthesiology 1974; 41:243).
In summary, general anesthesia itself may predispose to mild oxygenation defects by causing the development of atelectasis and decreasing the functional residual capacity. In the absence of chronic lung disease or some other preexisting pulmonary disease, the effects of general anesthesia on gas exchange are mild and usually transient. Effects of Thoracic Surgery on Pulmonary Function Abnormalities of chest wall function do not usually cause major postoperative respiratory dysfunction in cardiac patients because they have a median sternotomy for their surgical incision. 16 A thoracotomy results
in more lung trauma and lung compression and a greater need for postoperative analgesia than a sternotomy incision. l7 •18 Cardiac surgical patients that have internal mammary arteries harvested for their coronary artery bypass operation may have more pulmonary problems postoperatively because they may develop pleural effusions secondary to pleural trauma or hemorrhage. 19 These effusions can be large ~d can impair postoperative pulmonary function. MJr this reason, patients who have internal mammary arteries harvested usually receive a chest tube, whereas patients undergoing median sternotomies require only a mediastinal chest tube for postoperative drainage of blood. Sternal wound infections occur in approximately 1 to 2 percent ofpatients, although frank wound dehiscence and Hail chest are uncommon. 20 Postoperative atelectasis is common in patients after cardiac surgery; however, the precise etiology of the atelectasis is unknown. Since the atelectasis often is worse in the left than the right lower lobe (Fig 2), previous investigators have speculated that the cause may be related to retraction of the left lower lobe postoperative gastric distention, or during surger~ transient paresis of the left hemidiaphragm secondary to surgical or hypothermic injury to the phrenic nerve. 21 •22 However, a recent prospective study of patients follOwing cardiac surgery failed to demonstrate significant abnormalities in phrenic nerve function in the vast majority of patients undergoing cardiac surgery, including those patients who developed significant atelectasis. 22 In this study, 98 percent of patients had postoperative atelectasis and 70 percent of the patients had atelectasis of both the right and left lungs. Yet, there were no abnormalities in the phrenic nerve conduction studies or ~tudies of diaphragm function in 39 of the 44 patients. In the five patients with abnormal phrenic conduction studies,
2.2 2 1.1
...
1.1
~ o u
en M
1.2
J
0.1
·M S FIGURE 2. Atelectasis scores for right (hatched bars) and left (open bars) lungs for preoperative and postoperative days 1 to 6 and on discharge. On the preoperative and discharge days and days 1, 2, and 3, n is equal to 57, while on days 4, 5, and 6, n is equal to 46, 18, and 8, respectively. Bars are ± SE (Reprinted with permission from Wilcox et ale Phrenic nerve function and its relationship to atelectasis after coronary artery bypass surgery. Chest 1988; 93:693-98).
u
~
0.' 0•• 0.2 0
Preop
2
3
4
5
6
Discharge
Day Number CHEST I 95 I 2 I FEBRUAR'(. 1989
425
only two of the patients required more than 48 h of mechanical ventilation. There was no relationship between the type of cardioplegia used and the presence of postoperative phrenic nerve dysfunction. Another recent study found unequivocal paresis of the left phrenic nerve in only 5 of 52 patients (10 percent) following coronary artery bypass surger~ 23 Discriminant analysis of intraoperative variables indicated that more severe atelectasis was associated with the follOwing variables: increased number of coronary artery bypass grafts, a longer operative and cardiopulmonary bypass time, violation of the pleural space, failure to use a right atrial drain or a cardiac insulating pad, and a lower body temperature. Thus, the etiology of postoperative atelectasis in these patients was considered to be multifactorial. Effects of Cardiopulmonary Bypass on Pulmonary Function Cardiopulmonary bypass has been shown to alter pulmonary function. Although most patients tolerate cardiopulmonary bypass well, a number of studies have established that cardiopulmonary bypass may result in both a systemic and pulmonary capillary leak syndrome associated with fever, leukocytosis, renal dysfunction, transient neurologic dysfunction, and a bleeding diathesis. 1 These multiorgan effects have been called the "post pump syndrome." The mechanism for this syndrome may be related to the exposure ofblood to nonendothelial surfaces causing subsequent complement activation, platelet aggregation, and effects on the 6brinolytic system. 24-26 In addition, pulmonary sequestration of neutrophils may occur from complement activation and from shear stress generated during cardiopulmonary bypass by blood pumps, air bubbles, tissue debris, and platelet aggregates. 26 Thus, both pulmonary and systemic microembolization may ~ur during cardiopulmonary bypass. It has been well established that the duration of cardiopulmonary bypass has some relationship to postoperative respiratory problems. 1.24.27 For example, the severity of interstitial pulmonary edema after cardiopulmonary bypass is proportional to the duration of cardiopulmonary bypass. 28 Frank acute lung injury with alveolar flooding and increased permeability pulmonary edema occurs more frequently when the cardiopulmonary bypass period exceeds 150 min. 1 Acute lung injury after bypass can occur for other reasons including blood transfusion reactions or lung injury secondary to an idiosyncratic protamine reaction.29-31 Another possible mechanism for the lung injury in these patients may be related to neutrophil activation with disintegration and release of toxic lysosomal granules along with proteolytic enzymes that may result in both endothelial and epithelial
inju~ However, our own experimental studies have demonstrated that many neutrophils can appear in the lung without causing lung injury.32 This is corroborated by a recent study that we completed in humans where large numbers of normal neutrophils were attracted to the air spaces of the lung by a potent chemoattractant (LTBJ without any evidence ofpulmonary epithelial injur~ 33 Clearly, more than just the presence of numerous polymorphonuclear white cells is necessary for acute lung injur~ Another possible explanation for acute lung injury is that cardiopulmonary bypass causes an inadequate blood supply to the alveolar epithelium, which in turn results in an insufficient release of surface-active material (surfactant) from type 2 alveolar epithelial cells. There is no direct proof to support this hypothesis, although more than 90 percent of patients who have had cardiopulmonary bypass have evidence of mild to moderate atelectasis bilaterally,21.22 and the atelectasis may be due to abnormalities in surfactant production. Experimentally, atelectasis develops when lungs are ventilated without PEEP and exposed to the atmosphere,34 two processes that may affect the function of surfactant. The cold temperature maintained during bypass may also accentuate abnormalities of surfactant production or function. 35 Another possible mechanism to explain acute lung injury is that the cardioplegic solution that is used during cardiopulmonary bypass may injure the lung because it contains a high concentration of potassium chloride (20 mEq/L). This solution returns to the right atrium and it may then passively enter the pulmonary microcirculation, unless it is removed by a right atrial cannula. In fact, one study showed that failure to use a right atrial sump during bypass was a factor that related to higher atelectasis scores. 23 Thus, it is possible that the cardioplegic solution is toxic to endothelial or epithelial cells, so that abnormal or insufficient surfactant is formed, which would then predispose to the development of atelectasis. In summary, there are multiple mechanisms which may account for the atelectasis and mild to moderate widening ofthe alveolar-arterial oxygen difference that occurs in almost all patients follOwing cardiopulmonary bypass. In most patients, the gas exchange abnormality cannot be explained by the development ofpulmonary edema or obvious bronchospasm. The most likely explanation is that the function or release of surfaceactive material is abnormal, probably secondary to pulmonary ischemia during cardiopulmonary bypass or other factors that interfere with epithelial cell function. It is possible that diffuse closure of small airways, secondary to the release of bronchoconstricting mediators, such as thromboxane, also may contribute to the gas exchange abnormalities. Future studies are needed to investigate these possibilities. Respiratory Management after Cardiac Surgery (Mattha): WIene,-Ktonish)
A FIGURE 3. A, Posteroanterior chest roentgenogram of a 68-year-old man two days prior to coronary artery bypass surgery. Lung fields are clear. B. Anteroposterior chest film taken one day after coronary artery bypass surgery. An endotracheal tube is in place. There is evidence of both right and left lower lohe atelectasis with blurring of the right and left hemidiaphragms. There also may be small pleural effusions. The PaO, was 65 mm Hg with an Flo, of 0.7 on mechanical ventilation with a tidal volume of 1.000 ml plus 5 cm H,O PEEP.
Respiratory Abnormalities Following Cardiac Surgery The preceding three sections have considered the independent effects of anesthesia, thoracic surgery, and cardiopulmonary bypass on pulmonary function. Patients who undergo cardiac surgery are subject to the combined effects of all these factors plus any preexistent effects from chronic lung disease. All of the studies that have examined the gas exchange abnormalities following cardiac surgery have documented a substantial increase in the alveolar-arterial oxygen difference following even uncomplicated cardiac surgery. The primary physiologic basis for the widened alveolar-arterial oxygen difference is rightto-left intrapulmonary shunting.16·37 There is some evidence for a contribution from ventilation-perfusion mismatch as a secondary cause for the oxygenation defect.16·37 As already discussed, the mechanism for right-to-Ieft intrapulmonary shunting and ventilationperfusion mismatch is primarily related to atelectasis (Fig 3). When severe oxygenation defects occur in patients following cardiac surgery, other problems such as noncardiogenic pulmonary edema should be considered. 28 •31 However, in most patients alveolar edema does not occur; rather mild interstitial edema with a minimal increase in extravascular lung water is more common. J8 A number of experimental studies have shown that interstitial edema alone has a minimal effect on gas exchange; alveolar edema is required to cause major abnormalities in oxygenation. 39 .4o
The patients oxygenation after cardiac surgery may be worsened by the administration ofanesthetic agents that decrease hypoxic pulmonary vasoconstriction. \1,4\ In addition, vasodilating agents, especially nitroprusside, also will inhibit hypoxic pulmonary vasoconstriction and cause a decrease in the arterial oxygen tension. 42 Overall, respiratory failure after cardiac surgery is not common. When respiratory failure does occur, however, there should be a systematic search for the physiologic basis of the patient's respiratory failure. Table 1 provides a summary of the various causes of Table 1- Causes of Respiratory Dysfunction after Cardiac Surgery Decreased central respiratory drive General anesthesia, narcotics Neurologic injury Decreased respiratory muscle function Residual effect of muscle relaxants Pain, chest drainage tubes Poor cardiac function Severe obesity Diaphragm dysfunction (unilateral or bilateral) Exa('t'rbation of COPO Increased airway resistance Worsened bronchitis Alveolar diseases Pulmonary edema Cardiogenic Noncardiogenic Atelectasis Pneumonia
CHEST I 95 I 2 I FEBRUARY, 1989
427
respiratory failure following cardiac surgery. Inadequate alveolar ventilation, manifested by hypercapnea and respiratory acidosis, may be related to (1) a decrease in central respiratory drive, secondary to pharmacologic agents or to neurologic defects, (2) a decrease in respiratory muscle function (such as diaphragm dysfunction), or (3) an exacerbation of chronic obstructive lung disease with an increase in airway resistance. On the other hand, an oxygenation defect, manifested by a right-to-Ieft intrapulmonary shunt, is usually explained by diffuse atelectasis or, less frequently, cardiogenic or noncardiogenic pulmonary edema.
Respiratory Management after Uncomplicated Cardiac Surgery In general, most patients are ventilated for a period of 12 to 18 h following open-heart surgery in order to establish a period of hemodynamic stability and to reduce the immediate work of breathing postoperatively.43--IS Numerous studies recommend the routine use of overnight mechanical ventilation for postoperative cardiac patients. However, some investigators have proposed criteria to guide extubation within a few hours of surgery and have proposed physiologic parameters to predict that the patient will be able to ventilate spontaneousl~ Bendixen and colleagues46 noted that follOwing cardiac surgery, the average patient could be weaned successfully from mechanical ventilation if the vital capacity was > 10 mllkg of body weight. Other investigators have used an assessment of mental alertness, muscle strength, hemodynamic stability, and adequacy of pulmonary gas exchange as the basis for extubation. Some studies have used the duration of cardiopulmonary bypass as criteria for extubation. In one stud~ for example, 18 of 22 patients (82 percent) were extubated within 10 h of returning to the intensive care unit if the total bypass time was less than., 100 min. 47 None of these patients required reintubation. Patients who had more than 100 min of cardiopulmonary bypass time required longer periods of ventilation, and only nine (41 percent) patients in this group could be extubated within 10 h oftheir cardiac surger~ These investigators did not find that a vital capacity or a maximum inspiratory pressure could predict which patients would tolerate extubation and ventilate spontaneousl~
Other studies have indicated that the majority of patients can be extubated safely within a few hours following cardiac surger~ For instance, one group of investigators reported that 90 percent of patients were successfully extubated follOwing coronary artery bypass surgery without any increase in pulmonary morbidit~ 48 In another study, 63 percent of patients were 428
successfully extubated within 5 h of arrival in the intensive care unit. 49 Very recently, 143 patients with good ventricular function were managed without difficulty in an intermediate care unit after extubation in the operating room following cardiac surgery. 50 A prospective trial ofearly vs late extubation following coronary artery bypass surgery was undertaken at our own institution. Thirty-eight patients were randomly assigned to either early extubation (2 h postbypass) or late extubation (18 h postbypass). Anesthesia techniques consisted of inhalational agents followed by reversal of muscle relaxants in the patients who were extubated earl~ There were no significant differences between the two groups in subsequent respiratory or hemodynamic stability. The investigators concluded that early endotracheal extubation following uncomplicated coronary artery bypass utilizing halothane anesthesia is safe and does not increase postoperative cardiac or pulmonary morbidity. 51 In fact, a recent study evaluating the physiologic basis for impairment of oxygenation following coronary artery bypass surgery reported that the level of intrapulmonary shunt and ventilation-perfusion inequality did not change, whether or not the patient was intubated or breathing spontaneously37 (Table 2). Therefore, maintaining prolonged mechanical ventilation does not usually improve the patients pulmonary function in the routine, uncomplicated postcardiac surgery patient. In summary, extubation within a few hours follOwing cardiac surgery is reasonable in patients who are completely stable hemodynamicall}; have evidence of good ventricular function and have received an anesthetic compatible with early extubation. However, overnight ventilation for 12 to 18 h is a more reasonable approach in patients who have evidence of decreased ventricular function or have moderate to severe chronic lung disease. Also, patients who have hemodynamic instability intraoperatively or in the immediate postoperative period should be mechanically ventilated until their hemodynamic abnormalities are stabilized. The risks of continuing mechanical ventilation for 16 to 24 h postoperatively are minimal, especially when weighed against the possible complications of premature extubation in patients with postoperative arrhythmias, a low cardiac output, decreased or an agitated mental status, or exacerbated chronic lung disease or pulmonary edema. In order to wean patients from mechanical ventilation, it is reasonable to use a T-piece of CPAP trial to assess their work of breathing (minute ventilation, tidal volume, respiratory rate), as well as the effect of spontaneous ventilation on their arterial blood gases and vital signs. If they tolerate a trial on T-piece or CPAP ventilation, then extubation is ordinarily well tolerated. Respiratory Management after Cardiac Surgery (Ma~
W,.".,.KronIsh)
Table 2-Cas Exchange in Patients after Cardiac Surgery During Mechanical Ventilation and Spontaneous Breathing· Minute Ventilation. Umin
Arterial Po•• mm Hg
Arterial Pco,. mmHg
Patient
MV
S8
MV
S8
MV
S8
MV
S8
MV
S8
MV
S8
1 2 3 4 5 6 7 8 9
90 83 153 81 113 141 88 80 84 101 28
91 81 138 78 89 90
35 35 38 41 34 38 35 53
46
7.49 7.49 7.48 7.44 7.48 7.51 7.48 7.36 7.37 7.46 .05
7.38 7.38 7.36 7.43 7.38 7.42 7.41 7.35 7.29 7.38 .05
14.3 12.6 12.1 5.6 10.3 10.1 10.9 7.1 10.7 10.4 2.8
6.0 6.6 5.0 3.7 5.6 6.3 7.6 7.5 7.5 6.2 1.3
31.9 33.0 34.2 37.2 43.7 30.0 40.8 42.2 45.0 37.6 5.5
38.2 38.0 25.6 39.9 31.1 29.7 46.3 35.0 48.3 36.9 7.5
17.9 26.1 19.1 19.3 11.9 13.6 13.8 27.7 16.4 17.9 6.3
21.4 22.8 23.0 21.9 17.5 21.3 20.5 18.7 21.0 19.9 2.4
Mean Standard deviation p Value MV vs S8
85
92 79 91 18 NS
46
52
43
44
47 43
53 56 48 5
43
39 6
Arterial pH
0.01
0.001
0.02
Dead Space. %
NS
Shunt. %
NS
·Adapted from Dantzker DR. et al. Gas exchange alterations associated with weaning from mechanical ventilation fi,lIowing coronary artery bypass grafting. Chest 1982; 82:674-77. (Reprinted with permission of the American College of Chest Physicians.)
Respiratory Management of Pulmonary Edema after Cardiac Surgery The most common cause of pulmonary edema following cardiac surgery is high pressure in the pulmonary microcirculation from elevated pressures in the left heart. The usual cause is poor left ventricular function, although there may be a specific valvular defect as well. The ventricular dysfunction may be due to an unsatisfactory surgical repair, extensive disease that could not be surgically repaired, or ongoing ischemia. All of these patients require pharmacologic treatment to improve their cardiac function. A combination of vasopressors and vasodilators are often used, depending on pulmonary and systemic hemodynamic measurements. 52 The pulmonary edema will gradually clear as the ventricular function improves. Mechanical ventilation is required to support these patients during this time period to preserve adequate oxygenation, to decrease myocardial oxygen demand by allowing the mechanical ventilator to assume the work of breathing, and to decrease preload to the heart. Patients who develop noncardiogenic pulmonary edema or ARDS following cardiopulmonary bypass and cardiac surgery may have severe respiratory failure that requires a longer period of mechanical ventilatory support 1 (Fig 4). Although the possibility of sepsis must be considered in patients who have had cardiac surgery, the source of the acute lung injury usually is related to the duration of cardiopulmonary bypass, a blood transfusion reaction, or a protamine reaction. I .30.31 The natural history of ARDS following cardiopulmonary bypass has not been systematically studied. However, our experience indicates that many
patients with ARDS following cardiac surgery have a more limited form of acute lung injury that appears to be restricted to pulmonary endothelial injury. The prognosis for recovery from ARDS follOWing cardiopulmonary bypass may be better than from ARDS that is associated with sepsis or gastric aspiration. More studies are needed to evaluate the prognosis of patients with ARDS following cardiopulmonary bypass along the lines of recently proposed guidelines for defining acute lung injury. 53
FIGt:RE 4. Anteroposterior chest radiograph taken six hours after cardiopulmonary bypass and three-vessel coronary artery bypass surgery. The pulmonary artery wedge pressure was 5 mm Hg and the protein concentration of the edema fluid suctioned from the endotracheal tube was 95 percent of the plasma protein. thus confirming the diagnosis of increased permeability pulmonary edema or ARDS. The Flo. was 1.0. with a PaO. of69 mm Hg and 8 cm H.O PEEP. CHEST I 95 I 2 I FEBRUARY, 1989
429
The goals of managing these patients are the same as in all patients who have ARDS. Respiratory support for these patients depends primarily on the use of PEEP to maintain adequate oxygenation; PEEP significantly improves oxygenation in patients with ARDS, by increasing functional residual capacity and therefore allowing ventilation with a lower FIo2 .54 By lowering the FIo2 , PEEP may reduce further lung injury from oxygen toxicity. In normal lungs, lung impairment from oxygen toxicity can develop after 40 hours of breathing 100 percent oxygen;55 however, no toxicity develops when the FIo2 is less than 50 percent. 56 Diseased lungs may be more susceptible to injury from oxygen toxicity;57 thus, the safe level of oxygen delivery in ARDS is not known. However, in one experimental study, dogs were exposed either to air or to an FIo2 of 0.5 continuously for eight days after mild lung injury caused by oleic acid. 58 There were no differences in the clinical course, in the lung water content, or in the lung histology for the dogs breathing air or for those breathing oxygen. In this model of moderate lung injury, an FIo2 of0.5 appeared to be safe. Therefore, FIo2 less than 0.6 may not cause further lung injury in patients with diseased lungs, even with continuous administration. 59 Other potential benefits of PEEP have not been demonstrated. PEEP does not decrease lung water content. 60 When given prophylactically, PEEP does not protect against development of ARDS.61 Therefore, PEEP does not treat lung injury, but it may be useful in protecting the lung from further injury by allowing a decrease in the FIo 2 to safer levels. To be specific, PEEP improves oxygenation by increasing the lung volume at end-expiration - the functional residual capacity-which keeps air spaces open that would otherwise become atelectatic. The multiple-gas technique has provided evidence that PEEP decreases intrapulmonary shunting by changing exchange units from shunt to normal V/Q ratio units. 62 The benefit of PEEP on oxygenation is tempered by its major disadvantage: a reduction of cardiac output; PEEP decreases cardiac output primarily by increasing intrathoracic pressure, which then decreases venous return. 63 ,64 There are other more complex effects of PEEP on cardiac physiology, including possible changes in left ventricular compliance or contractilit); but the primary mechanism for reduction in cardiac output with use of PEEP relates to a decrease in preload. 65 The best PEEP (ie, that level at which there is the most benefit and the least harm) has been sought by many investigators. The value of using a level of PEEP to achieve the highest pulmonary static compliance has not been confirmed. 66 Others have suggested using the maximal PEEP possible to minimize FIo267 or intrapulmonary shunt. 68 However, given the risk of 430
barotrauma and the depression of cardiac output associated with high levels ofPEEE some investigators recommend using the lowest level of PEEP necessary to lower the FIo2 below 0.60. 59 ,69 When cardiac output decreases secondary to low levels of PEE~ either volume or inotropic therapy are recommended to facilitate the use of a level of PEEP sufficient to lower the FIo2 to below 0.6, if possibl~. Oxygenation also may be improved by other measures. Increasing the mixed venous oxygen saturation by increasing the cardiac output will improve arterial oxygenation. 70 Paralysis or sedation ofagitated patients may decrease oxygen utilization especially if the patient is shivering it may also improve the efficiency of ventilation by relaxing the diaphragm and chest wall. 71 If pulmonary disease appears asymmetric on the chest roentgenogram, repositioning the patient so that the better lung is dependent may decrease rightto-left shunting and improve oxygenation.72 Careful management of hemodynamics is important. The pulmonary arterial wedge pressure should be maintained as low as possible without compromising cardiac output. An elevated pulmonary artery wedge pressure will amplify the amount of the edema fluid that collects in the lung if there is an increase in lung vascular permeability.40,42 Currently, all treatment is supportive. In particular, corticosteroids are not beneficial in ARDS.73 A number of investigators have reviewed in detail the pathogenesis and management of patients with ARDS.74,75 Management of Patients with Chronic Lung Disease after Cardiac Surgery
Preoperative assessment of pulmonary function is important, especially if the patient has a long history of smoking or pulmonary symptoms. These tests will help define whether the patient has restrictive lung disease or obstructive lung disease. Chronic obstructive pulmonary disease is the most common cause of preoperative pulmonary dysfunction in patients undergoing cardiac surgery. The degree of pulmonary dysfunction can be determined by these preoperative pulmonary function tests. A number of patients have mild COPD, which will not significantly influence their postoperative management. Patients who undergo cardiac surgery with moderate to severe obstructive lung disease (FEV 1
