Hypoxemia is defined as a pressure of oxygen in blood (paO2) of less than 60 mm Hg, or a hemaglobin O2 saturation of less than 90%. The differential diagnosis of pathologies causing post CS hypoxemia is:

• Alveolar hypoventilation.
• Ventilation / Perfusion (V/Q) mismatch caused by the following clinical pathologies
• Airway obstruction caused by ETT compromise, bronchospasm, or airway secretions
• Hydrostatic pulmonary edema
• Adult respiratory distress syndrome (ARDS)
• Pneumonia
• Atelectasis
• Pulmonary embolism
• Hemodynamically active drugs
• Post operative pulmonary hypertension
• Decreased mixed venous oxygen (MVO2)
• Anatomic right heart to left heart shunting of venous blood.

If the lung's alveolar ventilation (Va), as a whole, is too low, more O2 is removed from alveolar air by capillary blood than is replenished, resulting in a low alveolar pressure of O2 (pAO2). Since pAO2 is the driving pressure of O2 into capillary blood, this in turn results in a low paO2. Similarly, when Va is low, CO2 is not removed from the alveoli and paCO2 rises above the normal value of 40 mm Hg. So, the first step in evaluating a patient who is suspected of being hypoxemic or known to be so from O2 saturation data is to obtain an ABG. If indeed the paO2 is low and the paCO2 is high, then the hypoxemia is very likely to be, at least in part, due to alveolar hypoventilation. Marked increase of pulmonary "dead space" can also increase paCO2, but this circumstance is very unusual, and discussed below in the section on Hypercarbic Rrespiratory Failure. Whether or not the hypoxemia is solely due to alveolar hypoventilation can be determined by utilizing the relationship between pAO2 and paCO2 in the "alveolar gas equation", which states:

pAO2 = (PIO2 - paCO2) / R - F

where PIO2 is the pressure of O2 in inspired air, R is the respiratory quotient (number of moles of CO2 produced per mole of O2 consumed), and F is a small correction factor. PIO2 can be derived from the FiO2, barometric pressure (pB), and water vapor pressure (pH2O). At sea level the FiO2 of room air is 0.21, pB is 760 mm Hg and pH2O is 47 mm Hg (at 37oC). Therefore PIO2 = 760 - 47 x FiO2 = 713 x 0.21 =150 mm Hg. R is almost always 0.8 under most physiologic conditions and F is so small that it can be ignored. So, on room air, at sea level, pAO2 = 150 - paCO2/0.8. Normally the pAO2 should be within 7 mm Hg + 0.27 x patient's age of the PaO2. If the pAO2 - the paO2 (alveolar-arterial oxygen difference (A-a O2D) is greater than this, then there is one of the other major pathophysiologic causes of hypoxia at play as well, specifically V/Q mismatch, anatomic right heart to left heart shunting of venous blood, or impairment of O2 diffusion across the alveolar capillary membrane.

For example, if a patient's O2 saturation is 85% on room air, the first step in the evaluation is to obtain an ABG. If the ABG shows a paO2 of 50 mm Hg, and paCO2 of 80 mm Hg, you know immediately that this patient's hypoxemia is very likely to be, at least in part, due to alveolar hypoventilation. To determine if there are other pathophysiologies at play, the alveolar gas equation is used to calculate this patient's pAO2. 150 - 80/0.8 = 150 - 100 = 50. A-a O2D is therefore 0, and alveolar hypoventilation is the sole cause of the hypoxemia. < /STRONG >

The A-a O2D can be calculated when patients are on different FiO2s, but there are two caveats to doing so. First, when external O2 is applied to a patient it is difficult to know the exact FiO2 the patient is breathing because the patient will almost certainly entrain some room air with inspiration. Secondly, as the FiO2 increases so too does the anticipated A-a O2D (43). Specific clinical entities causing alveolar hypoventilation are discussed in the section below on hypercarbic respiratory failure. Under normal physiologic conditions, a patient's pulmonary capillary flow (perfusion), is matched, or goes to, alveoli with good ventilation. When a significant amount of perfusion goes to relatively poorly ventilated alveoli (V/Q mismatch) the result is hypoxemia. Most post cardiac surgery hypoxemia is caused by this pathophysiology.

Airway obstruction results in poorly ventilated lung units and hypoxemia. Obstruction in post CS patients is commonly caused by ETT malposition (discussed above under Ventilator adjustment and trouble shooting), "plugging" of airways by mucous or mucopurulent secretions, and bronchospasm. Patients with the chronic bronchitis variant of COPD are at particular risk for airway plugging. These patients, as a baseline, produce large quantities of mucous, and often, because of impaired pulmonary mechanics, have a poor cough. Anesthesia, analgesia, and the pain of coughing associated with sternotomy often precipitate the build up of obstructing mucous in chronic bronchitis post cardiac surgery. Post operative patients who develop acute bronchitis with its attendant large quantities of mucopurulent secretions, for the same reasons are at risk for airway compromise.

Bronchospasm usually occurs in patients with underlying obstructive lung disease, either asthma or COPD, and may be triggered by a variety of perioperative factors. Among those factors are bronchial edema associated with the volume overload attendant to CPB, irritation of the airway by the ETT and suctioning, aspiration of gastric contents at the time of anesthesia, acute bronchitis, and very rarely an adverse medication reaction from beta blockers or narcotics, which can cause histamine release. Chest auscultation findings of patients with significant mucous or mucopurulent secretions include rhonchi and expiratory crackles, as well as focally decreased breath sounds in areas where bronchi have been completely occluded by secretions. Patients with bronchospasm demonstrate expiratory and occasionally inspiratory wheezes.

Therapy for airway secretions includes mechanical suctioning through an ETT if the patient is intubated or nasotracheal suctioning if the patient is not. Mucolytics are helpful, delivered either through nebulized acetylcysteine (which, it is important to realize, can prompt bronchospasm), or enterally using guaifenesin, most commonly (see Drug Table 3 at the end of the chapter). We avoid vigorous chest percussive therapy, at least in the immediate postoperative period, for fear of disrupting the sternotomy. Patients are encouraged to cough, however, while bracing their sternotomy with a specially designed pillow.

Bronchodilators help, as well, in promoting cough and allowing secretions to be mobilized. Of course, if acute bronchitis is at play we treat with antibiotics. For bronchospasm, the cornerstone of treatment is inhaled bronchodilators, which basically are of the beta 2 - adrenergic agonist or anticholinergic drugs. The most common beta - agonist is albuterol in either standard form or more recently developed (R) isomer form (levalbuterol). Levalbuterol has been shown to have a longer half-life, to be a more effective bronchodilator and have less cardiac effects (44). The only anticholinergic currently available is ipatropium bromide. We typically use both albuterol and ipatropium bromide to maximize bronchodilation. We avoid intravenous catecholamines and theophylline in managing bronchospasm because of the cardiac side effects. Drug table 1 (at end of chapter) lists various regimens for bronchodilating drugs. Bronchitis with the bronchospasm are treated with antibiotics. Patients with hydrostatic pulmonary or bronchial edema may also benefit from diuretics.

Many patients post-cardiac surgery, have an element of hydrostatic pulmonary edema (Conti 45). Hydrostatic pulmonary edema in the post-cardiac surgery patient is caused by a combination of the large volumes of fluid given to the patient to maintain CPB, the drop in oncotic pressure of the blood caused by the hemodilution of blood necessary for cardiopulmonary bypass, and sometimes cardiac dysfunction caused by either temporary stunning of the myocardium in CPB or congestive heart failure (46). Hydrostatic pulmonary edema is characterized by physical exam findings of end-inspiratory lung crackles especially at the lung bases, CXR demonstration of pulmonary venous congestion and a prominent interstitial pattern, A-a O2D elevation and decreased pulmonary compliance. Treatment of hydrostatic pulmonary edema is diuresis. This is the most effectively accomplished by standard loop diuretics.

Fully developed ARDS is a relatively rare occurrence following CPB but is associated with a mortality of 40 - 60% (15). ARDS can be caused by systemic inflammation prompted by CPB machinery membranes, aspiration of gastric contents at the time of anesthesia induction and, in the instance of multiple blood product transfusions perioperatively, transfusion related acute lung injury (TRALI).

The exact mechanism for the development of systemic inflammation associated with CPB is not completely clear. However, it appears to involve complement activation, which prompts neutrophil recruitment to the lung. These activated neutrophils, through release of toxic granular contents, are thought to be the cause of "pump lung" (15). It does appear that, though frank ARDS occurs in only 1 - 2% of CPB cases, a lesser version of lung injury associated with this inflammation is far more common, occurring in 12% of patient in a recent series (47).

Aspiration of gastric contents is also rare, but still does occur and can cause ARDS. Usually the anesthesiologist will have noted this occurrence at the time of induction, though not always. TRALI is generally caused by antibodies in the donor blood to the patient's native white blood cells, resulting in clumping of white cells in the lung, the so-called leukoagglutinin reaction. These clumped white cells are activated and produce inflammation of the lung and ARDS. Steroids appear to be helpful for this pathology in our and written experience (48). Very rare is a reaction to protamine which is given at the end of CPB to reverse heparin, involving an anaphylactoid mechanism, which can involve the lung in ARDS (49).

ARDS is clinically characterized by end-inspiratory crackles on lung auscultation, chest x-ray evidence of a diffuse interstitial and alveolar filling pattern, decreased lung compliance, and an elevated A-a O2D. Specific treatment of the acute phase ARDS is essentially supportive care. This means maintenance of mechanical ventilation until hopefully the ARDS resolves. It is very important, however, to properly set the mechanical ventilator. Recent studies have clearly shown the advantage of ventilator strategies which avoid of "overdistention" and "underdistention" of the lung (50). This is because in this, and many other types of lung injuries, there are lung units that are minimally affected and lung units which are severely involved. Vt will typically go to the less affected and more compliant lung units. Traditional sized Vt of 10-15 mL/kg can overdistend these lung units and injure them creating a spiral of worsening ARDS. Recent studies using a Vt of between 5 and 10 mL/kg, and not allowing PPs to exceed 30 cm of H2O pressure have shown significant improvements in mortality and other outcomes of patients with ARDS (51, 52) Part of the pathology of ARDS is injury to the type II alveolar cells, which produce surfactant. Without surfactant, many areas of the lung collapse. These "underdistended" lung units, are continually being opened by traditional volume or pressure cycled mechanical breaths only to collapse again upon expiration. This continual opening and closing creates destructive sheer forces to the alveolar membrane, and escalates lung injury, just as alveolar overdistension does. PEEP, originally developed to recruit collapsed lung units for ventilation to improve oxygenation, appears to be the most effective tool for prevention of this form of ventilator toxicity. The dose of PEEP is best judged utilizing volume pressure curves and inflection points (51). Discussion of these techniques is beyond the scope of this text. However, for practical purposes at the bedside, titrating PEEP in the traditional fashion for O2 improvement will result in a dose that has good effect in preventing "underdistension". We recommend starting with a PEEP of 3 cm. and go up in increments of 2 cm, not exceeding a PEEP of 15 cm. Follow arterial oxygenation to see that it is improving with increments in PEEP, as well as lung compliance. It is important to carefully follow the patient's blood pressure, as eventually, by increasing PEEP, cardiac output will become compromised. The dose of PEEP where oxygenation and compliance is best and where cardiac output is not compromised, is selected. This is usually between 5 and 12 cm of PEEP.

ARDS is now known to have a biphasic natural history. After the acute "inflammatory" phase the lung will enter either a resolution and reparative phase, or what is called now the "fibroproliferative" phase. This phase is marked by the influx of fibrosis producing inflammatory cells and destruction of the lung's capillary bed. Clinically, the fibroproliferative phase of ARDS is characterized by low-grade temperatures, progressive and severe decline in pulmonary compliance, a progressive increase in dead space, and gradually worsening CXR. This is accompanied by a need for increased minute ventilation to maintain adequate gas exchange. This phase of ARDS usually occurs three to seven days into the course of the disease, and there is now good evidence that steroids are a very effective therapeutic tool for fibroproliferative phase ARDS (53), and in our hands, has proved to be quite helpful. It is very important to make sure that there is no active infection, however, before instituting steroids.

Pneumonia and atelectasis provoke hypoxemia by causing effected lung units to be poorly ventilated. Pneumonia is discussed below (Bronchitis and Pneumonia). Atelectasis in CS patients is expected, to some degree, as part of the normal postoperative course (see Alterations of Pulmonary Physiology Following Cardiac Surgery). Additional problems and pathologies contributing to, or causing atelectasis include phrenic nerve dysfunction, pleural effusions, and splinting from postoperative pain. Physical examination is characterized by again end-inspiratory crackles, though these crackles typically are higher pitched, or "dry". CXR demonstrates low lung volumes and sometimes, obvious effusions, segmental or lobar lung collapse, or "plate-like atelectasis". Treatment of postoperative atelectasis consists, of encouraging cough, incentive spirometry, mobilization, and adequate pain control. We avoid the use of positive airway pressure devices such as IPPB, or IPV out of concern for disrupting the sternotomy. For patients on the ventilator, maneuvers to recruit alveoli consist of increasing the tidal volume, while of course watching the PIP, and utilizing PEEP in the same fashion as would be utilized for ARDS but at generally much lower levels.

Pulmonary embolism used to be thought of as a rare occurrence in the CS patient. However, the incidence in the largest series to date was 3.2% (18). Interestingly, all events occurred after the first three postoperative days. This is presumably because of the coagulopathy created by CPB. Important risk factors for the development of PE were prolonged recovery from surgery, obesity, and heparin induced thrombocytopenia. To prophylaxis against thromboembolic events it is our practice to use graded compression stockings for all CS patients in the immediate postoperative period, but out of concern for bleeding we do not use heparin until after the third postoperative day. Suspicion for pulmonary embolism, should occur when there is an abrupt worsening of oxygenation without an obvious cause. Physical examination findings include tachypnea, a split P2 upon cardiac auscultation, and jugular venous distension. CXR will often be clear, or unchanged, but occasionally demonstrate oligemia in one area of the lung, or a peripheral wedge shaped opacification signifying an infarct. A-a O2D will be elevated. Echocardiogram can be very helpful in assessing for the elevated pulmonary artery pressures and right ventricular strain that accompany large or multiple emboli. Diagnosis of PE usually requires a ventilation-perfusion scan, CAT scan of the lung utilizing special imaging techniques, or pulmonary arteriogram. Therapy of pulmonary embolism is anticoagulation, usually with heparin. Rarely, if a pulmonary embolism is massive and proximal in the pulmonary artery, the patient will require mechanical disruption or removal of the embolism by catheter techniques or surgical thrombectomy. In the immediate postoperative period lytic agents, such as TPA, are contraindicated. Hemodynamically active drugs have effects on the pulmonary vascular bed as well the systemic circulation. By thus altering perfusion patterns in the lung, V/Q mismatch can be created with resultant hypoxemia. Therefore, it is wise to monitor for adverse oxygenation effect while using vasoactive drugs, especially vasodilators. Postoperative pulmonary hypertension (fully discussed below in Postoperative Pulmonary Hypertension) causes hypoxemia by causing V/Q mismatch, predominantly, but also occasionally by prompting opening of the foramen ovale.

Decreased mixed venous oxygen occurs in patients who have decreased O2 delivery or markedly increased peripheral oxygen utilization. For practical purposes, this happens in post cardiac surgery patients who have impaired cardiac output and/or a high metabolic rate (usually from a major infection). Under these circumstances peripheral tissue oxygen extraction markedly increases resulting in a low mixed venous (pulmonary arterial) O2. This blood does not have adequate time to be fully oxygenated by the alveolar capillary unit, and so results in pulmonary venous blood that hypoxemic. Diagnosis of this pathophysiology as a cause of or contributor to hypoxemia is made presumptively under the right clinical circumstances and through mixed venous sampling. This is done most effectively through a pulmonary artery catheter. Therapy for this circumstance involves either increasing O2 delivery or decreasing peripheral O2 consumption. Increasing O2 delivery usually involves improving cardiac output through pharmacological or mechanical augmentation, but also improving O2 carrying capacity by increasing intravascular volume and hemoglobin, when necessary. Decreasing O2 consumption usually involves gaining control of infections and decreasing body temperature through acetaminophen and mechanical cooling devices.

Shunting of venous blood from the right side of the heart to the left heart without passing through the capillary bed for oxygenation requires both an anatomic conduit from right to left, and elevated pulmonary artery pressures. Surgery for CHD would, of course, be the most common setting for this to occur. Otherwise, as mentioned above, if pulmonary artery pressure rises severely or abruptly, an anatomically closed foramen ovale can become patent. Diagnosis of right to left shunting is best made with a contrast enhanced Echo Doppler study. Therapy for shunt begins with maneuvers designed to reduce pulmonary hypertension (see Postoperative Pulmonary Hypertension). We have found nitric oxide to be particularly helpful in this regard. Additional temporizing medical maneuvers include using systemic vasopressors to increase left sided cardiac pressures to discourage right to left shunting. Further treatment depends upon the anatomy of the shunt and discussion of such is beyond to scope of this article.

Hypercarbic respiratory failure

Hypercarbic respiratory failure characterized by a pCO2 >45 mm Hg is almost always caused, in the post CS patient, by alveolar hypoventilation. The differential diagnosis of alveolar hypoventilation in the post CS patient is:

• Decreased central respiratory drive.
• Phrenic nerve injury.
• Mechanical disruption of the chest wall.
• Peripheral neuromuscular dysfunction.
• Airway obstruction.

Decreased central drive to breathe in the post-cardiac surgical patient is nearly always due to over sedation. Rarely a perioperative cerebrovascular accident (CVA) involves the repiratory control center in the medulla of the brainstem. Occult or untreated hypothyroidism can cause an inadequate central drive in response to respiratory load (54), and is worth checking for in cases of unexplained hypercarbia. A small percentage of patients with sleep apnea will chronically retain CO2. These patients are usually obese, and have some obstructive lung disease as well ("obesity hypoventilation syndrome").

Phrenic nerve injury, as mentioned above, is fairly common post CS especially when those surgeries involve CPB with direct cold cardioplegia. A recent, very sophisticated study involving electrophysiologic testing of post-cardiac surgical patients for phrenic nerve dysfunction found it to be present in 26% of patients (55). Eight-five percent of the time, the phrenic nerve abnormality was on the left. All of the remainder of the patients except one had right phrenic nerve abnormalities. One patient had bilateral phrenic nerve abnormalities. In this study, patients were followed for up to one year post-cardiac surgery. In all patients, the phrenic nerve abnormalities completely resolved, usually within the first several weeks of the post-operative period. Unilateral phrenic nerve palsy should by itself not be enough to cause hypercapnic respiratory failure as we know from other studies that it impairs vital capacity by only approximately 25% (56). However, if unilateral phrenic nerve palsy and consequent unilateral diaphragm dysfunction is coupled with other pathologies, such as moderate-to-severe chronic obstructive pulmonary disease, it can certainly produce hypercapnic respiratory failure. Bilateral phrenic nerve palsy would be expected to produce hypercapnic respiratory failure. Though this is rare, it should be looked for in the post-cardiac surgical patient with otherwise unexplained hypercapnia. Phrenic nerve dysfunction is most easily evaluated by the "sniff test". This test involves having the patient sniff while imaging the diaphragms by fluoroscopy or ultrasound. Lack of diaphragm motion or paradoxical diaphragm motion, meaning the diaphragm moves cephalad rather than caudally, suggests phrenic nerve dysfunction. There is no specific treatment for phrenic nerve dysfunction other than good supportive care and otherwise unloading the ventilatory system from other burdens such as airway obstruction or secretions.

Mechanical chest wall disruption in the post-cardiac surgical patient for practical purposes means a sternotomy that has become unstable causing the patient to exhibit flail chest physiology. Clinically, this is characterized by retraction of the sternum during inspiration. Often, a flailed chest is accompanied by pulmonary secretion build-up as the patient's cough is very compromised. Therapy for flail chest, of course, involves surgical correction.

Peripheral neuromuscular dysfunction can rarely complicate the post-cardiac surgical course by impairing ventilator muscle function. This can occur when there are severe abnormalities of certain electrolytes, which are necessary for normal muscular function, to include potassium, magnesium, phosphorus and calcium. As mentioned above, rarely post-cardiac surgery, patients will develop systemic inflammatory response syndrome (SIRS). This syndrome can involve peripheral nerves and muscles in what has been called "neuropathy and myopathy of critical illness" (57). There is no specific treatment for these pathologies and they are very difficult to diagnose in the post-cardiac surgical patient. A common cause of hypercapnic respiratory failure in post-cardiac surgical patients is airway obstruction. For patients with advanced chronic obstructive pulmonary disease, the increased airway resistance attendant to CPB can be enough of an additional respiratory load to prompt respiratory failure. Patients with chronic bronchitis often have difficulty clearing their secretions secondary to impaired cough due to the pain of sternotomy. On physical examination, these patients will often be seen using accessory respiratory muscles and demonstrate decreased breath sounds suggesting airflow obstruction. Wheezes and rhonchi will often be auscultated suggesting bronchospasm and airway secretions. Therapy for these patients involves inhaled bronchodilators, the beta-adrenergic agonist and anticholinergic classes as well as mucolytics (see Drug Table 1 and 3 at the end of chapter). Endotracheal or nasotracheal suctioning is often required to clear airway secretions. We avoid vigorous chest physiotherapy and devices like IPPB and IPV, out of concern for disrupting the sternotomy.

Indications to place extubated patients back on the ventilator include hypercarbia with a pH of less than 7.3 and impending respiratory failure secondary to reparatory muscle fatigue. Some helpful clinical signs of respiratory muscle fatigue are: rapid shallow breathing (RR greater than 35 with small Vts), use of accessory respiratory muscles, and "abdominal paradox" (abdomen goes in rather than distends with inspiration). Sometimes reintubation can be avoided through the use of noninvasive positive pressure ventilation, and it is often worth a trial. We have had good experience with BiPAP applied through a nasal mask. Patients must, however, be alert and able to coordinate with the machine.

As mentioned above, hypercarbia is almost always caused by alveolar hypoventilation. Occasionally, however, a tremendous increase in physiologic "dead space" can cause hypercarbia. Dead space occurs when areas of the lung have more ventilation than they do perfusion. When dead space becomes very, very large, there is not enough capillary bed in contact with the ventilated lung to adequately release blood CO2. This occurs clinically in the circumstance of massive pulmonary emboli, advanced adult respiratory distress, and severe pulmonary hypertension. Diagnosis and therapy of these entities are discussed in the above section dealing with hypoxemia.

Pleural effusions

 Pleural effusions are common after CS. This is especially so after CABG, where the reported incidence is 40 to 90% in the immediate post operative period, depending upon the series (12,13). In most, though not all reports, bypass grafting utilizing internal mammary arteries (IMAs), rather than just saphenous venous grafts, has a significantly higher incidence of effusions (58). Heart failure or volume overload can certainly cause pleural effusions, but most instances of effusion in the immediate post operative period are not caused by such (13), and their exact etiology is unclear. Proposed mechanisms for the effusions include: a variant of the post cardiac injury syndrome where a local, pleural immunologic reaction produces the effusion, disruption of lymphatic drainage of the pleural space, especially when IMAs are harvested, leakage of mediastinal fluid through pleurotomy (often created in the process of IMA grafting) (14), or a reaction to topical hypothermia (59).

Most effusions are small, left sided, and of no clinical consequence (14). However, up to 10% of effusions are moderate to large and require treatment, for symptom relief (60). Early postoperative effusions are bloody exudates and eventually resolve after one or two thoracenteses, though they may persist in minimal quantity for several months. Rarely, effusions continue to recur after several thoracenteses. These effusions are exudative and typically contain lymphocytes as the predominant cell type (61). These patients can be tried on nonsteroidal anti-inflammatory agents and/or glucocoticoids (typically prednisone). In our experience, pleurodesis, best achieved through thoracoscopy is often required to permanently control the effusion. Two other mechanisms for the development of post CS pleural effusions are worth noting. The post cardiac injury syndrome can occur after myocardial infarction, cardiac trauma or surgery (62). It is characterized by fever, positional and pleuritic chest pain, pericardial and pleural effusions, often pulmonary and parenchymal infiltrates, and leukocytosis. The syndrome occurs in 10 - 40% of cardiac surgery patients, depending upon the series. Time to onset is one to several weeks after surgery with three weeks as an average. Antimyocardial antibody can be helpful in diagnosis though not definitive. Treatment is NSAIDs and/or glucocorticoids. There have been rare reports of chylothorax after cardiac surgery (63). These presumably result from trauma to the thoracic duct. Concern for chylothorax should be raised when pleural fluid is milky in appearance. Diagnosis is established by high levels of triglycerides in the pleural fluid. In most instances, with TPN and time, the chylothorax will resolve. Rarely, corrective surgery or pleuroperitoneal shunting will be required.

Bronchitis and Pneumonia

The incidence of postoperative pneumonia varies between 3 and 22% (64,65). The most recent series, using sensitive and specific diagnostic tools such as the protected microbe collection brush, place the incidence at between 5 and 10% (17). Occurrence of this nosocomial pneumonia peaks on postoperative day four (66), though we and others have seen well-developed cases on the first postoperative day. Interestingly, there is essentially no literature specifically addressing bronchitis in post cardiac surgery patients. This, presumably, is because of the difficulty of making a certain diagnosis of bronchitis in these patients. However, in our experience it is also a prevalent problem, and if not addressed leads to pneumonia. Pneumonia in post cardiac surgical patients is associated with a higher mortality and prolonged ICU and hospital stays (17).

These data, and our experience, have led to practices at our institution aimed at decreasing the incidence of postoperative pulmonary infections. We pay special attention to, and augment pulmonary toilet in patients with COPD through the use of bronchodilators, mucolytics, incentive spirometry, cough encouragement (or frequent suctioning if on mechanical ventilation), and adequate pain control while avoiding over sedation. Whenever practical, we keep the head of our patient's bed up at least 15 degrees and avoid NG tubes. We use sucralfate, again when practical, when stress gastritis prophylaxis is indicated. Additionally, it is helpful to understand that CS patients routinely experience dysfunctional swallowing for approximately eight hours post extubation (73). A prolonged, severe dysfunction is seen in 4% of the patients. It is usually associated with a very high incidence of pulmonary aspiration and pneumonia (74). Therefore, our protocols keep patients NPO for eight hours after extubation, and when oral intake is begun our staff is trained to monitor for aspiration. Several novel approaches to prevention of pneumonia in the post cardiac surgery patient population are being explored. Some institutions have begun using an ETT which comes equipped with a special catheter for continual suctioning of secretions above the ETT cuff (75). At our institution we have shown that chlorhexadine gluconate oral rinses significantly decrease the incidence of oropharyngeal colonization with nosocomial pathogens (manuscript in submission).

The classic clinical presentation of pneumonia usually includes fever, localized inspiratory crackles upon auscultation, focal infiltrate on CXR, leukocytosis, and production of purulent appearing sputum, the gram stain of which demonstrates many neutrophils and few epithelial cells, and a dominant organism. Unfortunately, there are other factors confounding the diagnosis of pneumonia in the postoperative patient. Pulmonary edema or ARDS can make physical examination and CXR findings difficult to interpret. Fever and leukocytosis are common in the immediate post operative periods as well. In these circumstances we have found sputum gram stain and quantitative sputum cultures to be invaluable.

Presentation of bronchitis also includes fever, purulent sputum, expiratory crackles, and rhonchi. Nosocomial pneumonia, in general, carries with it a mortality of 30 to 50%, but multiple studies have shown that empiric antibiotic therapy significantly lessens these numbers (71). Therefore, when pneumonia or bronchitis is clearly present, or strongly suspected, we culture the sputum and begin empiric antibiotics. Recent series have shown that the predominant organisms responsible for pneumonia in cardiac surgical patients are gram-negative enteric bacteria (17). Staphylococcus Aureus, however, has become a major pathogen in recent years accounting for nearly 19% of infections in a recent study (17). Empiric antimicrobial therapy, therefore, usually consists of a third or fourth generation cephalosporin and perhaps vancomycin. It is highly advised, though, to know the spectrum and common pulmonary nosocomial pathogens and their sensitivities at your institution to guide your empiric antibiotic choice. When sputum culture and sensitivities data become available it is very important to narrow the spectrum of antimicrobials as much as possible. Treatment involves, as well, application of good pulmonary toilet involving frequent coughing or suctioning, mucolytics, and bronchodilation if bronchospasm is present.

Pulmonary hypertension is common after CS, as is depression of right ventricular function (76). These effects, for most patients, are mild, last only hours, and do not significantly effect clinical course or outcome (77). However, patients with severe biventricular failure, and in those undergoing repair of congenital heart disease, mitral valve replacement, orthotopic heart transplant, or implantation of left ventricular assist devices are at risk for the development of severe pulmonary hypertension, and consequent right ventricular decompensation and death (78).

Pulmonary hypertension post cardiac surgery appears to be mediated by pulmonary vascular endothelial injury and dysfunction, which prompts vasoconstriction of the pulmonary vascular bed (20). A variety of factors can promote such injury and dysfunction, prominent among which are: intra and post operative hypoxia, hypercarbia, acidosis, intrinsically produced or extrinsically administered catecholamines, blood vessel wall shear stress associated with increased blood flow in conditions of left to right intracardiac shunt, and pathophysiologic changes sometimes attendant to CPB including pulmonary leukosequestration, production of inflammatory mediators, and, perhaps most powerfully, ischemia and reperfusion injury of the pulmonary vascular bed. Patients who go into CS with an already elevated pulmonary vascular tone, from, for example, advanced mitral valve disease, severe left ventricular dysfunction, or congenital heart disease, are particularly susceptible to these effects. When pulmonary vascular resistance markedly increases in these patients it can overburden an often already dysfunctional right ventrical leading to RV distension. This increases RV free wall tension and oxygen consumption as well as decreasing coronary artery perfusion. This sequence can spiral the RV into complete decompensation resulting in poor LV filling and cardiogenic shock.

Protamine, administered to reverse heparin effect at the end of CPB, can rarely cause a tremendous, abrupt rise in pulmonary vascular tone, which can produce right ventricular failure (79). The exact mechanism of this idiosyncratic reaction is unclear but has been suggested to involve complement activation and /or cyclooxygenase products.

In patients at risk for the development of deleterious pulmonary hypertension, avoiding hypoxia, hypercarbia, high mean airway pressures and acidosis is very important in preventing or at least lessening, this complication. A number of other experimental strategies for preventing pulmonary hypertension are being explored in these patients to include use of ultrafiltration during CPB, "substrate enhancement" with fructose-l, 6 diphosphate (FDP), and the addition of antioxidants during CPB (20).

To effectively treat pulmonary hypertension, and prevent RV failure, it is critical to recognize it's development early. This, of course, requires vigilance for its appearance in patients at risk. Physical findings of significant pulmonary hypertension include hypotension, jugular venous distension, a loud split P2, a parasternal or subxyphoid heave, an S3 or S4 heart sound which may vary in intensity with inspiration, hepatic distension and peripheral edema. The CXR will often show clear lung fields and an enlarged cardiac silhouette. Hypoxemia is caused by V/Q mismatching. A pulmonary artery catheter can be an invaluable diagnostic tool in this setting. It will show elevated pulmonary artery pressure as well as elevated right ventricular diastolic and right atrial pressures. In the setting of primary RV dysfunction without pulmonary hypertension, the pulmonary artery pressures will not be elevated. Echocardiography with Doppler can additionally be extremely helpful, as well, by demonstration of elevated PAP and RV pressure and volume overload, often concurrently with an underloaded LV.