CASE 5—2011Acute Respiratory Distress Syndrome in an Infant After Repair of Tetralogy of Fallot

ACUTE RESPIRATORY DISTRESS SYNDROME (ARDS) carries a high mortality. Protective lung ventilation (low tidal volume [TV] and positive end-expiratory pressure [PEEP]), particularly low TV ventilation, is the only therapy that has been shown to decrease mortality in ARDS patients (from approximately 40% to 30%).¹ ARDS and protective lung ventilation are usually associated with hypoxemia, hypercapnea, and increased pulmonary vascular resistance (PVR). These pathophysiologic changes can result in right ventricular (RV) dysfunction or aggravation of existing RV dysfunction. The authors discuss management issues of tetralogy of Fallot (TOF) in an infant who developed ARDS postoperatively after intracardiac repair.

Case Report⁎

A 12-month-old infant weighing 8 kg underwent uneventful intracardiac repair for TOF. A transannular pericardial patch was inserted for reconstruction of the RV outflow tract and main pulmonary artery. The patient had mild preoperative RV dysfunction (ie, mild hepatomegaly and full jugular veins) and was being treated with digoxin and furosemide. On the 3rd postoperative day, after an episode of vomiting, the patient developed breathlessness, bilateral chest crepitations and wheezing and became drowsy. The arterial oxygen saturation (SpO₂) decreased to approximately 85%. The patient was electively intubated and mechanically ventilated in volume-control mode with a TV of 55 mL (7 mL/kg), an inspired oxygen concentration (F_IO₂) of 0.8, a respiratory rate of 30 breaths/min, and PEEP of 5 cmH₂O. The monitored peak inspiratory airway pressure (PIP) was 39 cmH₂O, the heart rate was 166 beats/min, and the arterial blood pressure and central venous pressure were 76/35 mmHg and 8 mmHg, respectively. To support the circulation, intravenous infusions of epinephrine (0.05 μg/kg/min) and dobutamine (5 μg/kg/min) were started. The chest radiograph after intubation (Fig 1) showed diffuse bilateral infiltrates. The arterial blood gas (ABG) analysis showed a pH of 7.01, PaO₂ of 114 mmHg (PaO₂/F_IO₂ ratio = 142), PaCO₂ of 84 mmHg, and a base excess of 9.3 mEq/L. The F_IO₂ was then decreased to 0.6. Gradually, the PIP decreased to 32 cmH₂O and the ABG normalized.

Fig 1.

A chest radiograph after intubation revealing diffuse bilateral infiltrates.

During mechanical ventilation, total fluids (including intravenous fluids, parenteral nutrition, and Ryle's tube feedings) were limited to 100 mL/kg/d. Packed red blood cells were infused to maintain the hemoglobin at approximately 12 g/dL. After stabilization of the BP (approximately 90/50 mmHg) and the heart rate (approximately 130 beats/min), the intravenous inotropes were tapered and discontinued. After 2 days of mechanical ventilation, the lung infiltrates on chest radiography resolved (Fig 2). The ABG analysis at an F_IO₂ of 0.4, with a TV of 55 mL, a respiratory rate of 30 breaths/min, and a PEEP of 5 cmH₂O revealed a PaO₂ of 118 mmHg and a PaCO₂ of 44 mmHg. The airway secretions were minimal, the child was alert and active, the hemodynamic parameters were normal and stable, and the PIP had decreased to 21 cmH₂O. Thus, the patient was weaned from ventilatory support using an F_IO₂ of 0.4, pressure support of 14 cmH₂O, and a PEEP of 5 cmH₂O and was extubated. During weaning, the respiratory rate and TV were 28 breaths/min and 42 mL, respectively. The postextubation respiratory rate was 32 breaths/min, respiratory excursions were normal, and the ABG analysis on supplemental oxygen was as follows: pH of 7.40, PaO₂ of 272 mmHg, and PaCO₂ of 47 mmHg. There was no suggestion of airway obstruction.

Fig 2.

A chest radiograph after 2 days of mechanical ventilation revealing the resolution of lung infiltrates.

During the following 6 hours, the patient developed increasing tachypnea and hemodynamic instability. The ABG showed severe hypoxemia and respiratory acidosis (pH = 7.22, PaO₂ = 51 mmHg, PaCO₂ = 68 mmHg, and base excess = −0.7 mEq/L). The child was reintubated and ventilated in the volume-control mode at an F_IO₂ of 1.0 with a respiratory rate of 35 breaths/min, a TV of 50 mL, and a PEEP of 5 cmH₂O (PIP = 34 cmH₂O). Chest radiography showed reappearance of infiltrates and areas of collapse (Fig 3). The ABG showed a pH of 7.06, a PaO₂ of 54 mmHg, a PaCO₂ of 96 mmHg, and a base excess of −3 mEq/L. The PEEP and the TV were increased to 8 cmH₂O and 55 mL, respectively. The ABG analysis after 2 hours showed a pH of 7.21, a PaO₂ of 132 mmHg, a PaCO₂ of 74 mmHg, and a base excess of −1.6 mEq/L. The PIP had decreased to 27 cmH₂O, and the PEEP was decreased to 5 cmH₂O. Over 24 hours, the PaO₂ increased to 176 mmHg and the PaCO₂ normalized. Chest radiography showed a significant resolution of infiltrates and an expansion of lung fields after 4 days of ventilation. In view of the clinical improvement, the resolution of chest infiltrates, and satisfying weaning criteria, the patient was weaned from ventilatory support over 4 hours at an F_IO₂ of 0.5 and a pressure support of 14 cmH₂O. During weaning, the PEEP was decreased to 2 cmH₂O. Repeat chest radiography after extubation showed well-expanded lung fields (Fig 4). Postextubation, the child maintained satisfactory ABG on oxygen supplementation, and his later course in the hospital was uneventful.

Fig 3.

A chest radiograph revealing the reappearance of infiltrates and the areas of lung collapse.

Fig 4.

A chest radiograph revealing well-expanded lung fields.

Discussion

A sizeable number of patients develop RV failure after intracardiac repair of TOF. The pathophysiologic mechanisms of RV failure include inadequate myocardial protection, surgical ventriculotomy, residual RV outflow tract gradient, increased PVR, excessive airway pressure, volume loading, and/or acute pulmonary regurgitation.² The ARDS and ventilation strategies for ARDS (protective lung ventilation) potentially can aggravate RV failure because of increased RV afterload and pulmonary regurgitation secondary to hypoxemia, hypercapnea, increased PVR, PEEP, and increased airway pressure. The management of RV failure includes the manipulation of preload, afterload, myocardial contractility, and mechanical ventilation. Additionally, after intracardiac repair of TOF, the RV performance has been linked to RV perfusion. In an experimental study, Klima et al³ showed improved RV function by increasing the systemic arterial pressure. Therefore, RV afterload-reducing agents (eg, nitroglycerin, milrinone, and others) were not considered as options in this patient. Also, vasodilators are known to aggravate hypoxemia by increasing the pulmonary shunt. Similarly, considering the possibility of tissue congestion and the aggravation of RV failure with volume resuscitation, the RV preload in this patient was maintained in the modest range (CVP approximately 8 mmHg), and the total fluid administered was limited to 100 mL/kg/d. The authors also relied on clinical manipulation of myocardial contractility and mechanical ventilation parameters.

Low TV ventilation invariably results in hypercapnia, and PEEP might result in increased RV afterload; both hypercapnea and increased RV afterload have detrimental hemodynamic consequences. An important question is what should be the acceptable level of PEEP. The beneficial effects of PEEP include lung recruitment, the maintenance of recruitment after a recruitment maneuver, the restoration of functional residual capacity, the optimization of ventilation-perfusion, and an improvement in arterial oxygenation.⁴ Additionally, PEEP prevents expiratory collapse of alveoli and the shear stress that may occur because of the repeated opening and closing of alveoli and appears to modify the lung parenchymal injury induced by high transpulmonary pressure.⁵ Improved oxygenation and increased lung volumes decrease PVR. However, excessive PEEP can adversely affect pulmonary blood flow distribution and RV afterload. In a multicenter trial, Brower et al⁶ showed no difference in mortality with a higher PEEP compared with a lower PEEP (range of PEEP 2-8 cmH₂O). Gattinoni et al⁷ provided direct visual evidence, revealing that the capacity of “recruiting” collapsed lung units varies greatly among patients with ARDS, and in many patients an increase in PEEP cannot produce much recruitment because there is very little “recruitable” lung. In such patients, the use of PEEP may cause greater ventilator-induced lung injury by leading to pulmonary overdistention and hence actually may worsen the clinical outcome. In an editorial, Slutsky and Leonard⁸ stated that the “dose response” of PEEP may be such that most of the benefit is gained from increasing the PEEP from 0 to about 8 cmH₂O, and any further beneficial effects of PEEP in terms of decreasing lung injury caused by lung recruitment and derecruitment may be outweighed by the detrimental effects of PEEP (hemodynamic consequences or overdistention of the lungs). A decrease in cardiac output and the resultant decrease in oxygen delivery secondary to high PEEP can jeopardize organ perfusion, oxygenation, and function. To counter hemodynamic consequences of high PEEP and to optimize cardiac output, fluid resuscitation and vasoactive agents usually are used. However, aggressive fluid therapy in the presence of impaired RV function can precipitate RV failure, increase tissue congestion, and may further increase diffusion distance for oxygen and substrate delivery and jeopardize organ function.⁹ Arguably, the optimal level of PEEP, in the presence of ARDS and RV failure, is not easy to determine, and the best PEEP level would depend on the extent of the lung injury and RV failure. Considering beneficial as well as adverse hemodynamic effects of high PEEP and mandatory necessity of fluid resuscitation if high levels of PEEP were applied, the authors selected low levels of PEEP (2-8 cmH₂O) that would provide acceptable arterial oxygenation (PaO₂ >70 mmHg) without recourse to aggressive fluid therapy. The clinical features in this patient resolved over 2 days of mechanical ventilation. After the resolution of clinical features, this patient was weaned from ventilatory support using pressure support and a PEEP of 14 and 5 cmH₂O, respectively, and then tracheally extubated. However, within 6 hours of extubation, the patient developed extensive lung collapse and infiltrates and required reinitiation of ventilatory support.

The incidence of postextubation atelectasis in infants is as high as 35%.¹⁰ Extensive lung collapse after extubation suggests derecruitment caused by a loss of distending pressure (PEEP), a loss of surfactant and its function, and/or an accumulation of airway secretions leading to bronchial obstruction and lung collapse because air beyond the obstruction is absorbed. Endotracheal intubation further complicates this problem by impairing mucociliary clearance and by inhibiting an effective cough. Moreover, because of their small airway size and a less effective cough secondary to muscle weakness, small children are exquisitely sensitive to the obstructing effects of accumulating airway secretions. In an experimental study, Gattinoni et al⁷ showed derecruitment of alveoli when PEEP was withdrawn. Conceivably, areas of lung collapse may occur in these patients when PEEP is removed. Hence, it may be prudent to discontinue or reduce the PEEP to physiologic levels (2-3 cmH₂O) at the time of weaning from ventilatory support. Therefore, the PEEP was kept at 2 cmH₂O in this patient during weaning after the 2nd reintubation and the authors ensured that at this level of PEEP there was no loss of lung volume as assessed by chest radiography.

To summarize, a patient with TOF and impaired RV function suffered ARDS after intracardiac repair. The management issues were the prevention of aggravation of impaired RV function and complications of ARDS. The patient tolerated low TV ventilation, modest PEEP, and hypercapnea despite impaired RV function and did not develop worsening RV failure. After extubation using standard criteria, the patient rapidly developed respiratory failure, showed the reappearance of infiltrates and areas of collapse, and needed reinitiation of mechanical ventilatory support. The patient was weaned successfully and extubated after 4 days of ventilatory support. During weaning, PEEP was kept at 2 cmH₂O and full expansion of the lungs was confirmed before extubation. In patients having a tendency to develop lung collapse (ARDS), during weaning from ventilatory support it is prudent to reduce the PEEP to physiologic levels, and complete lung expansion should be confirmed before extubation.

Commentary 1†

In the presented case, the authors report on the management of postoperative ARDS and RV dysfunction in a 12-month-old infant after the repair of TOF. The presented infant had complete repair of TOF with a transannular patch. Shortly after extubation, the patient had a presumed episode of aspiration leading to progressive hypoxemia and an ARDS picture, requiring reintubation. The management of ARDS in this context is difficult because of the conflicting strategies typically used for the management of ARDS versus the conventional management of right-heart dysfunction after repaired TOF. This case provides an ideal forum for reviewing the pathophysiology and immediate postoperative management of RV dysfunction in patients presenting for intracardiac surgical repair of TOF.

The term “TOF” describes the most common form of cyanotic congenital heart disease, consisting of a large malaligned ventricular septum defect with overriding of the aorta, various degrees of right ventricular outflow tract (RVOT) obstruction, and RV hypertrophy. The anatomic spectrum of TOF and, in particular, the location and degree of RVOT obstruction (subvalvular, valvular, supravalvular, peripheral, pulmonary stenosis [PS]) are large and have major implications on patient management, including timing and the type of planned surgery. Early RV dysfunction in the immediate postoperative period as well as late RV failure typically seen 20 to 30 years after surgical repair have been described.11, 12 The various components of TOF all contribute to the pathophysiologic understanding of RV dysfunction in these patients.

The biggest contributor to RV dysfunction in the immediate postoperative period likely relates to the location and extent of the RV infundibular incision. The large ventricular septum defect resulting in a mostly right-to-left shunt (associated cyanosis and hypoxia) may cause myocardial damage even preoperatively and contribute to RV dysfunction in the postoperative period. The patient presented here was 12 months of age at the time of his surgery. If detected earlier, many centers would probably opt to operate earlier in life (eg, 3-6 months of age) because prolonged significant RVOT obstruction with RV pressure overload, persistent cyanosis, and hypoxia can lead to myocardial damage and fibrosis, resulting in myocardial dysfunction.¹³ The RV hypertrophy seen in TOF predisposes patients to RV dysfunction for several reasons. Myocardial protection during surgical repair is complicated by the thickened myocardium. Additionally, the hypertrophied RV is less compliant. Thus, diastolic RV dysfunction, with both a clinical picture and echocardiographic evidence of restrictive physiology, is seen. Myocardial damage during surgery (eg, poor myocardial preservation) has been associated with an increased troponin T peak after repair in patients who developed restrictive RV failure² and occurred more frequently when a transannular patch was performed.14, 15 Patients with a restrictive-type RV dysfunction after repair for TOF require more inotropic support, higher doses of diuretics, and have a prolonged intensive care unit (ICU) stay during the immediate postoperative period.16, 17 The long-term prognosis in patients with restrictive RV physiology, however, may actually be better when compared with patients with progressive RV dilation.¹⁸ Coronary artery variants with large conal branches crossing the RVOT may be sacrificed during complete repair because of the need for an RV infundibular incision, further leading to RV ischemia and dysfunction. The hypertrophied RV is also more prone to ischemia and depends on a higher diastolic perfusion pressure compared with coronary perfusion of a normal, thin-walled RV that is perfused during both systole and diastole. The degree and location of RVOT obstruction and additional peripheral branch pulmonary artery stenosis further add to predisposing patients with TOF to postoperative RV dysfunction. There is no mention in the presented case about the severity and location of RVOT obstruction. The surgeon inserted a transannular patch, presumably to relieve significant valvular PS. More and more, efforts are made to avoid a transannular patch if at all possible to prevent the serious complications seen with significant pulmonary regurgitation in the immediate postoperative period but even more importantly later in life. Mild postoperative residual PS even may be accepted in these patients for that reason. The decision to insert a transannular patch is based on preoperative echocardiographic criteria, yet ultimately rests with the surgeon exploring the degree of RVOT obstruction. If the size of the pulmonary annulus is prohibitively small, then a transannular patch will be required. If the pulmonic valve cannot be spared, free pulmonary insufficiency typically is seen upon discontinuation from cardiopulmonary bypass. This leads to acute RV volume overload and needs to be considered in the immediate postoperative management of these patients. Although significant pulmonary insufficiency can cause early RV dysfunction, patients typically recover quickly with appropriate management. The long-term consequences of significant pulmonary insufficiency only have been described in the last 10 to 20 years. Even significant pulmonary insufficiency is well tolerated for even a few decades, and usually symptoms of RV failure do not manifest until 20 to 30 years after the initial surgery.

In the presented case, acute respiratory distress developed in the infant soon after the removal of the endotracheal tube. In the setting of free pulmonary insufficiency, any increase in RV afterload (eg, increase in PVR) would be expected to increase end-diastolic RV volume as well as impose a pressure load on the RV, further impairing RV function. The presented arterial blood gas analysis showing severe acidosis and hypercapnia as well as ARDS and such associated pulmonary changes would indeed impose an increase in afterload on the RV. The authors followed commonly accepted recommendations for treating patients with ARDS and used lung-protective positive-pressure ventilation with low tidal volume and PEEP. The pathophysiology and treatment of RV dysfunction have been well described in detail elsewhere.19, 20, 21 A definition of RV failure has been suggested as “a clinical syndrome resulting from the right heart's inability to provide adequate blood flow to the pulmonary circulation at a normal central venous filling pressure.”²² The authors of this case report described measures to balance positive-pressure ventilation and PEEP-associated increase in RV afterload and pharmacologic circulatory support targeted to promote pulmonary and, subsequently, systemic blood flow. The authors chose mostly positive inotropic support (dobutamine and epinephrine) to accomplish this goal. In selecting the appropriate strategy for RV failure, the underlying specific congenital heart disease and associated structural findings, the specific etiology of acute exacerbation of RV dysfunction, clinical and laboratory findings, and response to therapy all should be considered. As discussed earlier, patients with TOF are prone to diastolic RV dysfunction presenting with echocardiographic features of restrictive filling patterns. In the restrictive-type RV failure, the stiff RV does not allow adequate filling or only at the cost of high filling pressures and acts more like a passive conduit. It would have been interesting to know the echocardiographic findings of RV failure in this patient. The authors only provide a central venous pressure reading of 8 mmHg, which is not elevated (would be expected to be significantly higher in a patient with restrictive RV physiology and/or free pulmonary insufficiency complicated by ARDS). The treatment for acute restrictive RV failure is targeted at maintaining left ventricular filling (thus systemic cardiac output) and perfusion pressure (with mostly passive blood flow through the RV) until RV function recovers. Positive inotropes such as epinephrine, dobutamine, and dopamine may be counterproductive and can worsen diastolic function and/or increase oxygen demand in an already struggling RV. Milrinone and, more recently, levosimendan have been suggested to be beneficial in this setting because of their lusitropic and vasodilatative properties in patients with restrictive RV failure.23, 24 In order to counteract the systemic hypoperfusion (from vasodilation) that may occur with the initiation of these drugs, the addition of a systemic vasopressor such as vasopressin is often useful and been shown to be beneficial and is a frequently used choice.²⁵ In cases of primarily systolic RV failure and/or increased PVR with a struggling RV, positive inotropic drugs are indicated, and in severe systolic RV failure epinephrine may be required to maintain transpulmonary blood flow. If an increase in PVR is the primary cause of RV failure, efforts to reduce PVR should be made. Selective pulmonary vasodilators such as inhaled nitric oxide, prostacyclin, and inhaled milrinone all have been used successfully.²⁶ Because of the route of administration (per inhalation), they exert fewer effects on the systemic circulation. Potent vasodilators have the potential to increase intrapulmonary shunt in patients with acute ARDS complicated by RV dysfunction. Drugs such as milrinone and levosimendan frequently are combined with positive inotropes and systemic vasoconstrictors for hemodynamic support.²⁷ Because RV failure commonly affects neonates after TOF repair, surgeons will often leave the patent foramen ovale open at the time of surgery, or, if an atrial septal defect is present, reduce its size, yet not close it completely. This allows for a “pop-off” of blood from the right to left atrium if the right atrial pressure increases because of right-heart failure. This will help preserve cardiac output at the expense of systemic arterial desaturation in the immediate postoperative period. This maneuver is extremely well tolerated in this group of patients because they were chronically cyanotic before the operation. For older patients undergoing TOF repair (greater than 3-6 months), this maneuver usually is not necessary unless there is concern about significant branch pulmonary artery stenosis because the incidence of primary RV failure is much less. In severe cases of RV failure in patients who do not have a native atrial-level communication and have low cardiac output despite aggressive medical management, it sometimes is necessary to emergently create an atrial-level communication in the cardiac catheterization laboratory with transseptal puncture and dilation of the atrial septum. In the presented case, the authors suspect that there was no atrial-level communication because arterial oxygenation was much higher than would have been expected if there was any right-to-left shunting.

In the presented case, there are insufficient data to diagnose the exact etiology of RV dysfunction. The low right-sided filling pressures and acute ARDS indicated a likely pulmonary cause, with an acute increase in PVR complicated by acute pulmonary insufficiency secondary to the transannular patch, rather than restrictive RV failure. The use of epinephrine and dobutamine certainly is indicated to support RV function in this setting, and the authors should be congratulated for the favorable outcome.

Commentary 2‡

This case highlights the importance of heart-lung interactions and how ventilatory strategies can support or impair RV performance. The authors discuss a case involving a 12-month-old patient recovering from surgical repair of TOF who suffered an acute lung injury presumed secondary to aspiration. The authors describe the successful management of this patient using a lung-protective strategy with low TVs and a moderate increase in PEEP. The initial arterial blood gas showed a mixed respiratory and metabolic acidosis that improved over 48 hours. Given that the traditional ventilator management in postoperative TOF patients has been to target a modest alkalosis with low mean airway pressures, the successful management of this patient is notable.

In this case, several factors lead to the belief that the right ventricle was recovering well after the operative repair (minimal cardiovascular support, readiness for extubation, low central venous pressures). After intracardiac repair of TOF, it is RV diastolic dysfunction that more often leads to low cardiac output. As in other congenital cardiac malformations, the maximal decrease in cardiac output typically occurs within 6 to 12 hours after separation from bypass.²⁸ As a supplement to the review by Mittnacht and Love, the present authors reviewed heart-lung interactions and suggested a few additional options that may enhance cardiac output in the setting of restrictive physiology.

A decrease in RV stroke volume occurs as the mean airway pressure increases and is the result of a decrease in venous return and an increase in afterload (pulmonary vascular resistance). Furthermore, relatively modest increases in pulmonary vascular resistance can lead to major changes in RV performance, even in the normal right ventricle.²⁹ Pulmonary vascular resistance is affected by the level of inspired oxygen, pH, airway pressure, and lung volumes.30, 31, 32, 33, 34 The clinical challenge is in determining the relative influence of these variables on pulmonary vascular resistance at any 1 point in time. Unfortunately, it is not an uncommon scenario to be at the bedside of a neonate or infant after the repair of TOF who is in a low-cardiac-output state because of restrictive physiology (ie, biventricular systolic function is usually well preserved in these patients, yet diastolic function may be markedly impaired). In this situation, there is antegrade diastolic pulmonary blood flow coincident with atrial systole. The right ventricle becomes unfillable during atrial systole and acts as a passive conduit. Additionally, there is simultaneous retrograde flow in the superior vena cava. These patients require higher filling pressures that usually result in pleural effusions and ascites, both of which compromise functional residual capacity and low lung volumes and require increasing positive pressure to maintain oxygenation and ventilation. These patients are also highly sensitive to changes in intrathoracic pressure, with forward pulmonary blood flow decreasing with increasing intrathoracic pressure. Therefore, positive-pressure ventilation contributes to low cardiac output in these patients. Interestingly, the use of negative-pressure ventilation has been shown to augment cardiac output in repaired TOF patients with restrictive physiology.³⁵ With positive-pressure ventilation, the goal is to limit inspiratory time, avoid prolonged plateau pressures, and minimize PEEP. Yet, with restrictive physiology, adequately oxygenating and ventilating the patient while limiting positive airway pressures can be nearly impossible.

As Mittnacht and Love nicely point out above, pharmacologic therapy with direct pulmonary vasodilators and lusitropic agents are appropriate in this setting and vasopressin aids in blood pressure support and may have less of a negative impact on diastolic performance than that of the catecholamines. In addition, as ventilatory requirements escalate, reopening the sternum or placing a peritoneal drain can restore functional residual capacity at lower airway pressures. High-frequency oscillatory ventilation may offer some advantages over conventional ventilation. Finally, extracorporeal mechanical support may be required until diastolic function recovers.³⁶ The silver lining in the postoperative presence of restrictive physiology is that pulmonary insufficiency is limited by the high atrial and right ventricular end-diastolic pressures, and this appears to favor better long-term outcomes.¹⁸ The authors agree with Mittnacht and Love in that there are insufficient data to determine whether the RV dysfunction is systolic or diastolic in nature and echocardiographic information would have been interesting.