FORMATION of noncardiogenic pulmonary edema has been observed after a variety of inciting events, including upper airway obstruction (negative pressure pulmonary edema [NPPE]),1acute lung injury,2anaphylaxis,3fluid maldistribution,4and severe central nervous system trauma (neurogenic pulmonary edema).5Both the diagnosis of pulmonary edema and an understanding of its underlying pathophysiology have important implications for treatment. Patients with severe postoperative noncardiogenic pulmonary edema who require mechanical ventilation should be ventilated with a low-tidal volume,6administration of positive end-expiratory pressure, and low plateau airway pressures.7,8Recent studies suggest that noninvasive respiratory support might be a viable approach for the treatment of patients with postoperative respiratory dysfunction, including postoperative NPPE.9
A 25-yr-old man (weight, 68 kg; height, 183 cm) presented to the surgery center for excision of back and thigh schwannomas on the same day. The patient's medical history was significant only for his history of multiple schwannoma resections and a history of smoking one pack of cigarettes per week for the past 5 yr. He denied previous problems with general anesthesia, and his baseline peripheral oxygen saturation was 99% on ambient air.
The patient was premedicated with 2 mg midazolam, and anesthesia was induced with 250 mg fentanyl, 500 mg thiopental, and 8 mg vecuronium given for facilitation of tracheal intubation. He was atraumatically intubated with a 7-mm ID endotracheal tube using a no. 3 Macintosh laryngoscope (Teleflex Medical, Research Triangle Park, NC) on the first attempt with direct visualization of the vocal cords. The patient was turned prone, bilateral breath sounds were reconfirmed, and schwannoma excisions were performed on the left thigh and the left flank. A total of 0.5 mg hydromorphone was administered for analgesia. The intraoperative course was unremarkable. The patient was hemodynamically stable with minimal blood loss and was easily ventilated and oxygenated. A total of 500 ml lactated Ringer's solution was administered during the 65-min surgical procedure. The patient was returned to the supine position for emergence and extubation. Nondepolarizing motor blockade was not reversed because train-of-four monitoring of the ulnar nerve showed a train-of-four ratio of greater than 90%, demonstrating adequate spontaneous recovery.
Immediately after extubation, the patient developed inspiratory stridor consistent with laryngospasm; the anesthesiologist had difficulty in mask ventilating the patient, and peripheral oxygen saturation decreased to less than 80%. Laryngospasm was treated by 50 mg propofol and manual positive pressure mask ventilation with 100% inspired oxygen. Peripheral oxygen saturation improved gradually, and the patient was transported to the postanesthesia care unit for further supportive treatment.
In the postanesthesia care unit, the patient's oxygen saturation was maintained with 100% oxygen administered via a nonrebreather facemask. The patient coughed pink, frothy sputum during the course of the first postoperative hour. Physical examination revealed crackles bilaterally at the lung bases, and a chest radiograph was performed, showing diffuse, bilateral, hazy, and interstitial opacity throughout both lungs, with normal lung volumes, normal heart size, and no pleural effusions (fig. 1). A diagnosis of NPPE was made, and the patient was admitted to the inpatient postoperative recovery room for overnight observation. With supplemental oxygen, diuretic treatment, and bronchodilator inhalation, his respiratory status continued to improve with peripheral oxygen saturations greater than 94% on ambient air 10 h after surgery. Examination on the morning of the first postoperative day revealed clear lungs bilaterally and peripheral oxygen saturation of 95–97% on ambient air. He was discharged later that morning without signs or symptoms of respiratory compromise on oral analgesics and usual surgical follow-up in 1–2 weeks.
Postoperative Recovery Room Diagnostic Evaluation and Treatment
A chest radiograph taken immediately after postanesthesia care unit admission showed diffuse bilateral opacities, a finding that was observed despite conservative intraoperative fluid management (fig. 1). The patient's history, operating room course, and clinical and radiologic findings were most consistent with pulmonary edema with NPPE as the likely cause; however, aspiration pneumonitis (Mendelsohn syndrome)10and diffuse alveolar hemorrhage resulting from upper airway obstruction11were also included in the differential diagnosis.
When considering the differential diagnosis of acute-onset perioperative pulmonary edema, both cardiac and noncardiac causes should be taken into account (table 1; fig. 2). Cardiogenic edema is usually preceded by new-onset left heart dysfunction and may be caused by acute ischemia, infarct, and/or severe arrhythmia, and the diagnosis is confirmed by echocardiography or measurement of the pulmonary artery occlusion pressure. It is likely that a combination of cardiogenic and noncardiogenic mechanisms contributes to the pathogenesis of postoperative pulmonary edema in many cases. For instance, although fluid overload itself can cause pulmonary edema in the presence of normal or even increased cardiac output,12intraoperative intravascular fluid overload can exacerbate chronic compensated heart failure.
Pulmonary edema caused by anaphylaxis is seen in the setting of exposure to a known or unknown allergen. In the perioperative setting, these often include neuromuscular blocking agents, antibiotics, anesthetics, or latex.13The onset is sudden and is typically accompanied by rash, urticaria, and swelling, but bronchospasm and hemodynamic collapse are frequently presenting symptoms. The clinical picture, time course, and severity, and its occurrence after administration of an allergen, help the clinician to relate signs and symptoms of pulmonary edema to an anaphylactic mechanism. The increased histamine and tryptase levels obtained immediately after the reaction are consistent with anaphylaxis. Radioallergosorbent tests and skin tests performed 4–6 weeks after a presumed reaction can help to confirm the clinical diagnosis and identify the inciting allergen.13
Neurogenic pulmonary edema typically occurs in the setting of a recent severe brain insult, such as subarachnoid hemorrhage, stroke, status epilepticus, trauma, or intracranial mass. Neurogenic pulmonary edema is typically accompanied by unregulated sympathetic discharge leading to pulmonary hypertension,14which induces stress failure of pulmonary capillaries and subsequent high permeability pulmonary edema.15
Acute respiratory distress syndrome and acute lung injury represent a heterogeneous group of severe hypoxic lung diseases. Activation of and damage to the pulmonary endothelium are the hallmark of acute lung injury or acute respiratory distress syndrome,16which is caused by a variety of inciting events such as sepsis, systemic inflammatory response syndrome, aspiration, caustic inhalation, blood transfusions, or trauma. Diagnosis is made by exclusion of other causes, as outlined in figure 2. The severity of hypoxic respiratory failure, chest radiographic findings, and the time course to recovery are key elements that need to be considered for making diagnosis of acute lung injury or acute respiratory distress syndrome. The edema fluid to plasma protein ratio is an additional method to discriminate between cardiogenic pulmonary edema and acute lung injury. Ware et al .2compared protein concentration (Biuret method) in the pulmonary edema fluid (taken via a suction catheter inserted into the endotracheal tube) and blood. Using a predefined cutoff of 0.65, the edema fluid to plasma protein ratio had a sensitivity of 81% and a specificity of 81% for the diagnosis of acute lung injury.
Before making the diagnosis of NPPE, other causes of pulmonary edema (table 2; fig. 2), particularly those requiring a rapid intervention (fluid maldistribution, anaphylaxis, and cardiogenic pulmonary edema), must be considered. In this patient, intraoperative fluid overload as a mechanism of pulmonary edema was not considered reasonable because the patient had only 500 ml isotonic solution administered intraoperatively, no history of left heart failure, and had been fasting overnight. There was no evidence of cardiogenic or neurogenic pathology and no signs or symptoms of anaphylaxis. Aspiration pneumonitis can be of increased concern in the prone position given the potential for increased abdominal pressure. Our patient was positioned on chest bolsters that allowed the abdomen to hang freely, which might help to decrease intraabdominal pressure. In addition, the radiologic picture of symmetric bilateral pulmonary interstitial infiltrates would be unusual for aspiration pneumonitis, which typically shows a localized infiltrate. In the immediate setting, we could not rule out acute lung injury or acute respiratory distress syndrome, but the severity of respiratory failure and the time course of clinical and radiologic recovery were not ultimately consistent with this etiology. Residual postoperative curarization is associated with reduced pharyngeal muscle tone and possible resulting upper airway obstruction.17In our patient, direct measurement of the train-of-four ratio by accelerometry showed a train-of-four ratio greater than 0.9, reflecting adequate recovery from muscle relaxant effects.18Coupling these considerations with the clinical picture of laryngospasm, we concluded that the patient's pulmonary edema was likely induced by negative intrathoracic pressure, potentially resulting from strong inspiratory efforts in the setting of laryngospasm.
In accordance with the reported data, symptoms and clinical signs of pulmonary edema resolved rapidly.19Although not performed in this patient, and typically unnecessary to make the diagnosis, hemodynamic measurements, including pulmonary artery occlusion pressure, pulmonary arterial pressure, and central venous pressure, taken after the development of edema, are typically normal.20
In this patient, conservative treatment with supplemental oxygen administered as 100% oxygen by a nonrebreather mask (flow, 15 l/min), 10 mg furosemide intravenously, and bronchodilators was started.21The patient's symptoms of pulmonary edema improved rapidly, such that noninvasive pressure support ventilation was not required. The rapid improvement of the patient's disease represents a typical case of acute postoperative pulmonary edema (table 2).
Postoperative NPPE typically occurs in response to an upper airway obstruction, where patients can generate high negative intrathoracic pressures, leading to a postrelease pulmonary edema. The current literature regarding its epidemiology is sparse. Young, healthy, athletic patients seem to be at risk for this disorder,22and the prevalence of postoperative NPPE is approximately 0.1%.22,23In patients developing acute postoperative upper airway obstruction, NPPE has been reported at an incidence of up to 11% (table 2).24
Typical events leading to acute upper airway obstruction accompanied by perioperative NPPE include laryngospasm and endotracheal tube occlusion by biting. Less typically, NPPE can also occur after foreign body aspiration, oropharyngeal surgery, or postoperative residual curarization,25which typically impairs the upper airway dilator muscle strength while preserving inspiratory muscle function.26Case reports and retrospective data suggest that the patient characteristics that increase the risk of NPPE seem to include younger patients in American Society of Anesthesiologists physical status categories I and II, who are thought to be most capable of generating highly negative intrathoracic pressures during an obstructing event. Procedural characteristics increasing the risk of NPPE may include oropharyngeal surgery (especially for tumors or other potentially obstructing masses) although the true incidence and hazard ratios have not been reported.23
Pathogenesis of Noncardiac Pulmonary Edema
Diagnosis of noncardiogenic pulmonary edema requires an understanding of the pulmonary fluid homeostasis. The Starling equation describes the equilibrium of fluid flow through a semipermeable membrane:
where Q = the net transvascular flow of fluid, K = the membrane permeability, P mv = hydrostatic pressure in the microvessels, P pmv = hydrostatic pressure in the perimicrovascular interstitium, πmv = plasma protein osmotic pressure in the peripheral vessels, and πpmv = protein osmotic pressure in the perimicrovascular interstitium.
The osmotic pressure is exerted by solutes in the blood versus those in the interstitium, which cannot cross the semipermeable membrane. Under normal conditions, most of this filtered fluid from the capillaries is returned to the systemic circulation by lymphatics.27The alveolar spaces, because of tight junctions in the alveolar epithelium, have very low permeability and do not fill with fluid. Disturbances of pulmonary fluid homeostasis can be induced by four pathways that can lead to increased interstitial fluid: increased hydrostatic pressure in the pulmonary capillary bed (or conversely, decreased pressure in the interstitium), decreased osmotic pressure of plasma, increased permeability of the membrane, and decreased return of fluid to the circulation via lymphatics.27,28
Pathogenesis of NPPE
During upper airway obstruction and forceful inspiration, pressure in the trachea and lower airways will decrease markedly. The pressure in the pleural space decreases by exactly the same amount, and the pressure in the pulmonary vessels decreases by much less, thus increasing the pressure difference between inside and outside the capillaries and accelerating the formation of interstitial fluid.
Two different mechanisms may explain the development of pulmonary edema during airway obstruction. The most likely mechanism relates to the observation that high negative intrathoracic pressures cause significant fluid shifts from the microvessels to the perimicrovascular interstitium, as seen in patients with congestive heart failure or fluid maldistribution states. The second proposed mechanism involves the disruption of the alveolar epithelium and pulmonary microvascular membranes from severe mechanical stress, leading to increased pulmonary capillary permeability and protein-rich pulmonary edema.
Evidence for a hydrostatic mechanism of NPPE comes from experimental and clinical data.29,30In an experimental model of NPPE, Loyd et al .29induced a negative inspiratory pressure in sheep (a 9 mmHg decrease in mean central airway pressure). Left atrial pressure decreased by 8 mmHg, and lung lymph flow was increased twice at baseline. Pulmonary arterial pressure was unchanged. The authors concluded that inspiratory loading is associated with an increase in the pulmonary transvascular hydrostatic gradient, possibly by causing a greater decrease in interstitial pressure than in microvascular pressure. Healthy human subjects can generate very high levels of negative inspiratory pressure (>100 mmHg), which in turn increases the return of blood to the right side of the heart, concomitantly increases pulmonary venous pressures, and decreases “downstream” pulmonary interstitial perivascular pressure. The negative intrathoracic pressures generated during the Mueller maneuver (inspiratory effort against a closed glottis) will result in an increased afterload,31which in turn will augment the pulmonary capillary hydrostatic pressures. Consequently, a marked increase in hydrostatic pulmonary pressure gradient can be generated, such that fluid filters out of the microcirculation and into the lung interstitium. When a critical quantity of edema fluid collects in the interstitial compartment, alveolar flooding occurs.32
Although many patients with NPPE recover with conservative management as in this case, some patients with severe NPPE (or underlying cardiopulmonary disease) require temporary intubation and mechanical ventilation with positive end-expiratory pressure.33Diuretics are often administered, but their use is controversial and may even be unnecessary.19
The patient's wheezing was thought to represent bronchoconstriction, which we treated with inhaled bronchodilators; however, wheezing is caused by air flow through narrowed airways, and this may not necessarily be due to bronchospasm. Turbulence within bronchi, irrespective of the cause, including interstitial edema induced narrowing of bronchial lumina, may account for the development of the clinical symptom wheezing. In vitro and in vivo studies in human and animal models show that β agonists may increase the rate of alveolar fluid clearance via increased active cation transport.34Although it is unclear how much nebulized salbutamol arrived at the alveolar epithelium in our patient, it is possible that bronchodilator administration may have accelerated regression of symptoms of pulmonary edema.
An alternative to intubation is noninvasive respiratory support (i.e ., noninvasive positive pressure ventilation or treatment with continuous positive airway pressure). Recent data suggest that noninvasive respiratory support may be an important tool to prevent or treat acute respiratory failure while avoiding intubation. The aims of noninvasive respiratory support in the context of NPPE include: to partially compensate for the affected respiratory function by reducing the work of breathing; to improve alveolar recruitment with better gas exchange; and to reduce left ventricular afterload, increasing cardiac output and improving hemodynamics.35Evidence suggests that noninvasive respiratory support may be an effective strategy to reduce intubation rates, intensive care unit and hospital lengths of stay, and morbidity and mortality in postoperative patients.9,35Ultimately, NPPE is a generally benign condition typically resulting in full recovery in 12–48 h when recognized early and necessary supportive treatment is instituted for hypoxemic and/or hypercapnic respiratory failure.
The immediate consequence of the Mueller maneuver is a markedly negative intrathoracic pressure, leading to increased pulmonary transvascular hydrostatic pressure and vulnerability to accumulation of filtered fluid in the interstitium and, ultimately, in the alveoli.
In addition to a hydrostatic mechanism of NPPE, there is evidence that the increased wall stress (circumferential wall tension caused by the transmural pressure) will alter the permeability coefficient (K ) of the endothelial barrier. A classic paper by John B. West, M.D., Ph.D., D.Sc. (Distinguished Professor of Medicine and Physiology, School of Medicine, University of California, San Diego, San Diego, California), et al .36studied the effects of increased capillary transmural pressure in isolated rabbit lungs. The number of breaks in the endothelium increased with perfusion pressures, suggesting that high capillary hydrostatic pressures cause major changes in the ultrastructure of the walls of the capillaries, leading to a high-permeability form of edema. This suggestion was subsequently translated into a human model of increased capillary transmural pressure. This study was performed in six healthy athletes 1 h after an extensive cycling exercise. Analysis of bronchoalveolar lavage in healthy athletes after cycling exercise revealed a higher erythrocyte count and increased protein and albumin content compared with controls, indicating disruption of the endothelial membrane and stress failure. This suggests that acute increases in transmural pressures such as in NPPE may lead to increased permeability of the endothelial barrier.37
Some information is available on the molecular mechanisms involved in increased endothelial barrier permeability in response to wall stress. When an acute increase in transmural pressure occurs, the radial expansion of the capillary wall translates into linear cellular stretch. Compared with shear stress from laminar flow, the response of endothelial cells to linear stretch is maladaptive.38,39Oxidative stress is one mechanism for injury that seems to be up-regulated by increased linear stretch. In fact, increasing levels of cyclic linear stretch result in up-regulation of inducible nitric oxide synthase40and xanthine oxidoreductase, as has been shown by Abdulnour et al .,41both of which have been repeatedly implicated in cellular injury and increased vascular permeability. Future studies will show whether these mechanisms of increased vascular permeability are clinically relevant in patients presenting with NPPE.
The authors thank Deborah Pederson, M.D. (Instructor in Anesthesia, Department of Anesthesia, Critical Care, and Pain Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts), for reviewing the case scenario and Fumito Ichinose, M.D. (Associate Professor of Anesthesia, Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Harvard Medical School), for his suggestions regarding the algorithm for making a diagnosis of negative pressure pulmonary edema.