Persistent Air Leak

There are three simple scenarios that describe persistent air leak: persistent air leak after parenchyma injury, after anatomic lung resection, and in ventilated patients. Persistent air leak after parenchyma injury can occur because of penetrating injury, blunt trauma with maceration or rib penetration, or in patients with underlying predisposing parenchyma lesions, primarily bullous disease. In this setting, management has followed the algorithm for spontaneous pneumothorax. Simple tube thoracostomy suffices in more than 80% of cases as long as there is full expansion. Occasionally, placing the chest drain to water seal will actually hasten resolution as the transpleural gradient is diminished. Air leak during more than 3 days or associated with recurrent pneumothorax appear to be most efficiently managed by thoracoscopic approaches than persistent chest drainage. Schermer and colleagues reviewed the course of 39 trauma patients who, except for air leak, were ready for discharge. This was determined by air leakage over more than 3 days. Twenty-five agreed to video-assisted thoracoscopic surgery (VATS) with reduced chest tube duration (total 8 vs. 12 days) and length of stay (10 vs. 17 days). Of course, technical factors should be ruled out (tube dislodgement or disconnection). Computed tomography (CT) scans can help define local lesions that may be amenable to thoracoscopic wedge resection or glue application, which may also prompt earlier VATS. In many instances, simply breaking down soft loculations and placing a chest drain under direct vision is the primary therapeutic benefit of thoracoscopy. We prefer to not use chemical pleurodesis, but rather pleural abrasion as we feel that this reduces the risk of parenchyma trapping and the uncertain long-term impact of chemical agents in younger patients. Patients with underlying lung lesions should be managed as they would in nontrauma circumstances. A final option in patients with prohibitive operative risks or small leaks is to convert the patients to Heimlich valve and manage them outpatient. As many as 80% will seal within 3 weeks using this approach.

As lobectomy and pneumonectomy are rarely performed for traumatic injury, it follows that the incidence of air leak (bronchopleural fistula [BPF]) is also small. However, the nature of acute lung resections is such that the risk is higher than after elective resection. Risk factors include long stumps, devascularization, and contaminated hemothorax. Ideally, after lobectomy/pneumonectomy, the stump should be reinforced either at the time of original resection or second-look exploration with pleural, intercostals, or other flap. Once it occurs, management is determined by timing (less than or more than 7 days postoperatively), degree (ventilatory compromise and whether the defect can be visualized endoscopically), physiologic status, and whether the patient is ventilated. BPF may present in stable patients as a new productive cough, with a drop in pleural fluid levels (after pneumonectomy) of two or more rib spaces, or new air–fluid level. In ventilated patients, empyema and loss of tidal volume may predominate. The primary goal is to prevent aspiration. In nonintubated patients, this is best accompanied by upright positioning or affected side down. Then, drainage should be instituted if a chest drain is not in place. If a drain is not in place, the new drain should be placed above the thoracotomy scar, as the diaphragm tends to rise to the level of the scar and adhere. If the leak is small, and endoscopically the hole cannot be clearly visualized, it is reasonable to attempt bronchoscopic glue application. Reoperation and stump closure are possible within 7 days, but the associated empyema increases the risk of failure. The longer the interval between the initial and second operation, the greater the difficulty. After pneumonectomy, the mediastinum becomes inflamed, the stump can only with difficulty be visualized, and mobilization is essentially impossible. Thus, after pneumonectomy, the best option is probably to occlude the stump with omentum, pack the chest with packs, and plan serial washouts until the leak scarifies closed. An alternative approach, particularly after right-sided pneumonectomy, is to perform transcarinal right mainstem resection. The residual stump cannot be removed as it tends to be fixed, but the mucosa should be cauterized and omentum or other viable tissue should be used to reinforce the new stump. The empyema cavity can then be treated by the drainage procedure of the surgeon’s choice. After lobectomy, similar options are possible, but further resection may be required (e.g., right middle lobectomy after right lower lobe stump leak).

Persistent air leak in a ventilated patient without a discrete lesion is better thought of as an alveolar-pleural fistula rather than a BPF. Clearly, the underlying lung injury affects outcome, with alveolar-pleural leak in adult respiratory distress syndrome patients being associated with up to 80% mortality. Whatever the underlying anatomy, air leak in ventilated patients can be a significant marker of increased mortality. Pierson and colleagues reviewed the course of 39 patients (out of a population of 1700 mechanically ventilated patients) who presented with air leak lasting more than 24 hours, of whom 27 were trauma patients. The risk factors for mortality correlated with the following: air leak not present on admission or shortly thereafter (45% early vs. 94% if developed later); leak greater than 500 ml per breath (57% if less vs. 100% if greater); and post–chest trauma (56% for trauma admissions vs. 92% for nontrauma admissions). These findings illustrate that while the course in trauma admissions is more benign, it still represents a major concern. On the other hand, the air leak itself is rarely the cause of death. These air leaks can lead to persistent or even tension pneumothorax that compromises ventilation. Pleural tubes (at times multiple) may be required. Less commonly, air leak is significant enough to affect oxygenation. The primary treatment is to minimize alveolar pressure, using end-inspiratory plateau pressure as an (admittedly crude) reflection of this. Ideally, the end-inspiratory plateau pressure should be less than 30 cm H₂O. The most common method of attaining this is to combine low tidal volume and permissive hypercapnia. Alternative methods if this approach fails are high-frequency jet ventilation or independent lung ventilation. It should be stressed that high-frequency jet ventilation, although used successfully in patients with central airway disruption and in the operating room, does not reduce mean airway pressure consistently, nor does it uniformly reduce air leak or improve oxygenation. Thus, it should not be used routinely in patients with alveolar-pleural fistula. A temporizing technique is to isolate the lobe that is the primary source of leak bronchoscopically. This is done by sequentially occluding bronchi with a Swan-Ganz or other balloon catheter. If this results in elimination or significant reduction in air leak, occlusive material (Gelfoam, fibrin glue, blood mixed with tetracycline, etc.) can be injected. In most cases, the air leak will diminish as airway pressure decreases. Surgery can be performed, but in the setting of diffuse parenchyma injury, lung inflammation, severe emphysema, and/or steroids, the risk is that staple lines will fail and the leak will be exacerbated. If surgery is felt to be needed, reinforced staple lines (i.e., with bovine strips, etc.), apical tents (mobilizing the apical pleura so that it falls onto the area of resection), and/or anatomic lobectomy (if predominantly one lobe) should be considered.

Pneumatocoele/Hematoma

Pneumatocoeles occur when disruption of lung parenchyma leads to internal rather than external leak of air and/or blood. They occur more commonly after blunt injury, but can be seen occasionally with deep stab or low-caliber missile injuries. These lesions are thus best described as a pulmonary laceration. They are usually solitary, at times multilobulated, and occasionally multiple. They are typically not apparent on initial radiographs, because small size and/or a superimposed contusion or hemorrhage obscures them. Over time, they evolve into thin-walled cavities with air and/or fluid. The location and size are affected by the mechanism. Compression leading to rupture, the most common mechanism, tends to be associated with central lesions. Compression, leading to shear forces, tends to present as an elongated paramediastinal cavity extending from hilum to diaphragm, and may be confused with loculated pneumothorax. Rib penetration forms tend to be small and peripheral. Adhesion tears are the least common. In the vast majority of cases, pneumatocoeles are benign. In rare cases, they may result in persistent air leak or become infected, in which case they are treated as abscesses.

Hematomas are formed by the same mechanisms that result in pneumatocoeles. They may remain solid, or with partial evacuation they can develop an air-fluid level or even a fibrin wall resulting in a crescent of air on the superior surface that mimics a fungus ball. Usually these lesions resolve over 3–6 months, and recognizing the shrinking process is one method to avoid confusing these with malignant processes.

Pneumonia

Pneumonia may be the most common complication of chest trauma. Risk factors include aspiration, need for ventilation, direct injury, pulmonary contusion, and persistent atelectasis secondary to pain. The incidence is as low as 6% in nonintubated patients to as high as 44% in ventilated patients. Despite the high incidence, there are no data supporting prophylactic antibiotics. Of all patients admitted with a diagnosis of pulmonary contusion, nearly 50% will develop pneumonia, barotrauma, and/or major atelectasis. One-fourth will progress to adult respiratory distress syndrome. Ventilatorassociated pneumonia (VAP), defined as pneumonia arising more than 48 hours after initiation of mechanical ventilation, is difficult to define and diagnose. At our institution, we have found that the incidence in patients ventilated for more than 7 days ranges from 20% to 30%. Clinical suspicion is often raised by new infiltrates, recurring fever, rising leukocytes, and/or a change in endotracheal secretions. However, distinguishing between colonization and infection may require specialized techniques. Quantitative cultures obtained from a variety of approaches increase the specificity (although perhaps with reduced sensitivity) of endobronchial cultures, and each institution must define cut-off values based on whether the patient is already receiving antibiotics (Table 1).

Table 1 Yields of Diagnostic Tests for Ventilator-Associated Pneumonia

Necrotizing Lung Infection

Necrotizing lung infections comprise a triad of clinical scenarios that overlap or can be present concomitantly. These are lung abscess, necrotizing pneumonia, and lung gangrene. All three are similar in that lack of perfusion is combined with tissue devitalization. In simplistic terms, lung abscess can be described as a region of necrosis less than a lobe with viable surrounding or bordering parenchyma. Lung gangrene represents complete lobar or entire lung destruction, often with only a rim of tissue remaining. Lung necrosis is best represented by patchy, often nonanatomic, loss of perfusion with variable parenchyma destruction, often seen on radiograph as multiple small abscess-like cavities. Although the three can be discussed separately, in most cases two or three coexist and so the management can also overlap.

The cause(s) of lung abscess in the surgical intensive care unit (ICU) population include aspiration, complications of pneumonia, retained foreign body, septic emboli, and/or infected traumatic injury. More specific etiologies in the trauma population include aspiration (with or without bronchial obstruction), infected pneumatocoele, infected site of resection (in particular emergent tractotomy), and late complications of ventilatorassociated pneumonia. As a whole, these are less common in trauma patients than nontrauma patients. Of 45 thoracotomies performed at our institution over 7 years for abscess, necrotizing pneumonia, and lung gangrene, only 4 were in patients initially admitted after traumatic injury.

The diagnosis of lung abscess may be relatively simple. Fever, purulent sputum production, or hemoptysis may prompt chest radiograph, which will identify an air-fluid cavity. On the other extreme, a persistently febrile patient in the ICU with dense consolidation may require CT scan before the underlying cavity can be recognized.

Over the three decades of approximately the 1950s through the 1970s, a number of advances reduced the mortality rates from approximately 50% to 10%. These advances included recognizing the importance of antibiotics, the role of aspiration, the need for pulmonary toilet (including liberal use of bronchoscopy), and finally the benefit in selected patients of operative intervention. Subsequently, the major addition to the armamentarium has been image-guided catheter drainage as an intermediate category between medical and surgical management. Percutaneous catheter drainage can be performed even in ventilated patients and has reduced the number of thoracotomies required. While there is always concern about the risk of empyema and/or bronchopleural fistula, the former can be usually easily managed by chest drainage, and the latter is rarely so significant as to impair oxygenation. Some patients will require thoracotomy, which, in the trauma population, usually results from persistent sepsis and inability or incomplete drainage, hemoptysis, or persistent or major bronchopleural fistula (see Table 1). The two primary operations are lobectomy for large central cavities, or debridement (plus muscle flap to help close the space) for smaller peripheral cavities. At operation, there are several technical points that can help reduce complications: prevent aspiration by isolating the affected lung before posterolateral positioning; expose the main pulmonary artery early in the case so that control can be achieved should hemorrhage result; place a nasogastric tube or esophagoscope in the esophagus because the anatomy may be obliterated; and refrain from resecting small abscesses (<2 cm) that are in otherwise viable parenchyma. Air leak is not uncommon, and, as will be discussed under the empyema section, a residual space can be managed with continuous postoperative irrigation.

The distinguishing characteristics of lung gangrene are central vascular thrombosis and bronchial obstruction, leading to significant cavitation and/or lobar or whole-lung liquefaction. As opposed to lung abscess, there is no firm, well-defined capsule. Both these features are defined by CT with intravenous contrast, and either one predicts the failure of medical therapy. This is because medical therapy relies on both blood supply for antibiotic therapy to be effective and on bronchial patency to allow expectoration of purulent material. Schamaun and colleagues followed 14 patients with unilateral complete lung gangrene. Four were treated medically and all died, while 10 underwent surgical resection with 100% survival. Some patients have diffuse bilateral disease. In the face of persistent signs of infection, if there is a primary target site, surgery is still possible and can be performed even if the patient cannot tolerate independent lung ventilation. Interestingly, dissection in the fissures and of the vessels is relatively easy, as the necrotic tissue tends to be easily swept aside. However, surgery resection should not be performed if the patient is pressor dependent. In this setting, it is better to temporize with pleuroscopy to treat associated empyema and percutaneous drainage of the large cavitary lesions.

Necrotizing pneumonia is characterized by areas of dense consolidation, patchy perfusion, and often multiple small cavitary changes. Percutaneous drainage does not help in this setting. Generally, parenchyma resection is not indicated. However, serial CT scans can identify areas that are developing demarcation lines, and in the setting of persistent pulmonary sepsis resection can be a reasonable option.