Surgery remains the primary therapy for patient with early stage non-small cell lung cancer. Nonsurgical modalities of local therapy such as focused stereotactic radiation and radiofrequency ablation have been used recently, particularly in patients thought to be of high operative risk. In an era of multimodality therapy, with the availability of several emerging treatment options for local therapy, it is imperative that we understand the potential complications of surgery for lung cancer as well as treatment strategies for these complications to select the best treatment options for individual patients.

The risks for pulmonary resection depend primarily on the preoperative status of the patient with regard to factors such as comorbidities, overall functional status, and preoperative pulmonary function. Complications of pulmonary resection that are well described and will be discussed in this chapter include pneumonia, respiratory failure, empyema, bronchopleural fistulae, prolonged air leaks, chylothorax, postpneumonectomy pulmonary edema, postpneumonectomy syndrome, and death.

Several contemporary studies have demonstrated improved outcomes in patients undergoing pulmonary resection for malignancies compared to resections performed in an earlier era. Recent improvements in patient survival and reduction in complications have been attributed to better patient selection for surgery as well as improved postoperative care, although these perceptions are not well supported by objective data.

Recently, the American College of Surgeons Oncology Group (ACOSOG) Z0030 trial randomized patients with non-small cell lung cancer undergoing an anatomic resection for clinical T1 to T2, NO to N1 tumors to mediastinal lymph node sampling or mediastinal lymph node dissection.¹ This study included 1023 patients from 63 different institutions and 102 different surgeons. There were 42 pneumonectomies, 42 bilobectomies, 766 lobectomies, 70 segmentectomies, and 101 patients with a combination of resections. The overall mortality rate in this patient group was 1.4%, and the overall morbidity rate was 38%. The most frequent morbidities were atrial arrhythmias (14%) followed by chest tube drainage >7 days (11%), air leak >7 days (8%), respiratory (7%), hemorrhage (3%), chylothorax (1%), and recurrent laryngeal nerve injury (1%). Table 39.1 outlines the incidence of significant complications occurring in both historical and contemporary series of lung cancer resections for comparison. The ACOSOG Z0030 multi-institutional study provides a contemporary benchmark of morbidity and mortality for patients undergoing major pulmonary resections for stage I and II lung cancers when performed by surgeons specializing in thoracic surgery.

In published series of major pulmonary resections for lung cancer, the reported mortality rates have ranged from 1.3% to 5.5% (Table 39.2).²,³,⁴,⁵,⁶,⁷,⁸ Wada et al.⁸ reported a series of 7099 thoracic procedures with significant differences in mortality noted depending on the extent of lung resection. The mortality rates were 3.2% for pneumonectomy, 1.2% for lobectomy, and 0.8% for sublobar resections. There were also significantly more deaths in patients with increasing age where the mortality rates were 0.4% for patients younger than 60 years, 1.3% for patients aged 60 to 69, 2.0% for patients 70 to 79, and 2.2% for patients 80 years or older. In the Z030 trial, there were no trends toward increased mortality or complications with larger pulmonary resections although the trial was not designed to examine differences in resection strategy.¹ Romano and Mark⁵ reported a direct correlation between the extent of pulmonary resection and mortality. In this retrospective study, the mortality rates were 3.8% for wedge resections, 3.7% for segmentectomy, 4.2% for lobectomy, and 11.6% for pneumonectomy. Factors that significantly correlated with patient mortality were patients older than 60 years, extent of resection, male gender, chronic lung or heart disease, diabetes, and volume of pulmonary cases performed at each participating institution. The finding that mortality rates are higher when cases are performed by general surgeons rather than surgeons
specializing in thoracic surgery has been suggested in other studies but remains controversial.⁶,⁷

Clinical stage of disease plays an important role in identifying the high-risk surgical patient for two reasons. The first is the perceived likelihood of achieving the goal of cure of cancer. Patients with early stage disease have improved survival over patients with advanced stage disease. As a rule, only patients with stage I to IIIA disease are even considered for surgical therapy. Within this group, patients with stage I disease have improved survival over those with stage IIIA disease. The surgical team may favor a more aggressive approach in a patient with earlier stage disease because of an increased likelihood of cure. The second reason clinical stage is important is because of the extent of resection needed to obtain tumor free margins. Tumors requiring chest wall resection, sleeve resection, vascular reconstruction, and pneumonectomy have higher rates of complications. At times, the risk of complications is high enough to avoid recommending surgery to less surgically fit candidates with more advanced tumors.

Although a basic understanding of the expected morbidity and mortality rates is important for discussions with patients before and after major lung resections, further risk stratification based on patient comorbidities and the extent of the planned operation are also necessary. Strand et al.⁷ reviewed the Norwegian experience of patients undergoing pulmonary resections between 1993 and 2005. During this period, there
were 4395 procedures with an overall mortality rate of 4.4%. Multivariate analysis identified several procedure-related factors associated with an increased risk of mortality including male gender (odds ratio [OR] = 1.76), age 70 to 79 or ≥80 (OR = 3.38 and 9.94 compared to patients ≤50), right-sided tumors (OR = 1.73), and bilobectomy or pneumonectomy (OR = 3.06 and 4.54, respectively). In this study, neither the extent of local-regional advancement nor tumor size had any impact on 30-day mortality. The presence of metastatic disease was the only tumor-related factor that had a negative impact on survival. The largest observational study evaluating the impact of patient’s age found a direct correlation between the patient’s age and mortality with the most significant increased risk occurring after the age of 70.⁷ The 30-day mortality OR for patients aged 70 to 79 and 80 to 89 were 4.91 and 19.71, respectively, when compared to patients ≤50 years. Most studies have shown that advanced age is a predictor of morbidity and mortality after pulmonary resection,⁷,⁸,⁹,¹⁰ although studies in carefully selected patients older than 80 years of age undergoing pulmonary resection have shown similar morbidity and mortality rates to younger patient cohorts.¹¹

In addition to factors associated with the extent of resection required, scoring systems exist to further stratify patients according to medical comorbidities.³ Algorithms for identifying higher risk patients include the Cardiopulmonary Risk Index (CPRI),¹² the Physiological and Operative Severity Score for Enumeration of Mortality and Morbidity (POSSUM),¹³ and the EVAD scoring system. The EVAD scoring system is primarily based on the patient’s forced expiratory volume in 1 second (FEV₁), age, and diffusing capacity of lung for carbon monoxide (DLCO).³ In a comparison of these three systems in a cohort of 400 patients, POSSUM scores that were retrospectively calculated did not correlate with either fatal or nonfatal complications of surgery.³ The CPRI scoring system did not correlate with patient mortality but did correlate with nonfatal complications when analyzed together. More specifically, the CPRI calculations correlated significantly with pulmonary and cardiopulmonary complications but did not correlate with cardiovascular, infectious, or other complications. The EVAD scoring system showed trends toward prediction of mortality and did correlate significantly with all types of morbidity evaluated with the exception of infectious complications. Unfortunately, the statistical predictive power of this scoring system is, at this point, not sufficient for application to individual patients and has not been further evaluated in other institutions. In addition, scoring systems specifically developed for lung resection, such as the Predictive Respiratory Quotient (PRQ)¹⁴ and the Predicted Postoperative Product (PPP),¹⁵ have not gained widespread use. Although the surgical team may not rely completely on a particular scoring system to determine the surgical fitness of a patient, many of the factors that make up these scoring systems along with surgical judgment are used to make the ultimate decision. Some of these factors may include clinical stage of disease, patient age, spirometry, and presence of comorbidities.

An important component of risk assessment is to determine whether there is sufficient pulmonary reserve to sustain the patient postoperatively. Pulmonary function testing is used to assess airflow, lung volume, lung mechanics, and gas exchange. The two most commonly used parameters for determining surgical resectability are FEV₁ and DLCO. Surgical risk has little or no correlation with preoperative FEV¹ when lung function is relatively preserved. An observational study of patients categorized into either normal lung function (FEV₁ >80% predicted, mean = 92.2%) or reduced lung function (FEV₁ ≤80% predicted, mean = 64.2%) found no differences in the rates of air leaks >7 days, atelectasis, bleeding, contralateral pneumothorax, or atrial fibrillation.¹⁶ However, the value of preoperative FEV₁ evaluation is more important with a greater reduction in pulmonary function. The risk of pulmonary resection increases substantially if the preoperative FEV₁ is less than 60% predicted.¹⁷ In general, patients who fall below this mark need further testing with ventilation/perfusion scanning, which is used to determine preoperative split lung function and predicted postoperative FEV₁. DLCO is used to evaluate the integrity of the alveolar capillary membrane for gas exchange. Ferguson et al.¹⁸ found in a retrospective study of 237 patients that preoperative DLCO predicted postoperative pulmonary complications in patients with acceptable FEV₁. They subsequently confirmed the importance of DLCO by showing that predicted postoperative DLCO is a predictor of morbidity and mortality independent of FEV₁.¹⁹ In addition, patients with DLCO less than 60% for pneumonectomy and less than 50% for lobectomy have an increased risk of postoperative complications.²⁰ Although these factors are frequently utilized in the preoperative risk assessment, it should be noted that others have not found spirometric parameters such as FEV₁ or DLCO to be significant predictors of postoperative complications.²¹,²² Others have successfully utilized spirometry in combination with exercise tolerance as a means of predicting pulmonary complications following resection.²³ Vital capacity, exercise induced hypoxemia (Δ PaO₂), low oxygen uptake/body weight (VO₂ max/BW), and Δ PaO₂/Δ VO₂ max/BW may be associated with an increased risk for complications.

After poor pulmonary reserve, the next most important factor contributing to risk of pulmonary resection is the presence of concomitant organ system dysfunction. In general, preoperative dysfunction of each of a particular organ system makes the risk of postoperative complications involving that organ system more likely. It is therefore essential to assess these organ systems preoperatively by history, physical exam, and special studies as needed.

The most common, immediately life-threatening intraoperative complications of pulmonary resection not related to anesthesia
management are massive hemorrhage, cardiac ischemia, arrhythmias, and contralateral pneumothorax. Common intraoperative complications that cause significant morbidity and mortality postoperatively are nerve injuries and injuries to the esophagus and thoracic duct.

Hemorrhage Massive intraoperative hemorrhage is usually the result of an injury to a pulmonary artery or vein branch sustained during dissection. The pulmonary artery and its branches are especially thin walled and easily injured during manipulation or traction employed to increase exposure. In contrast, the walls of the pulmonary vein are more resilient and withstand surgical manipulation much better. The risk of a difficult dissection and pulmonary artery injury can be anticipated in patients who have had induction chemotherapy or prior irradiation. In addition, patients with mediastinal granulomatosis or prior silica exposure will have regional bronchopulmonary lymph nodes densely adherent to branch pulmonary arteries. In such cases, it is prudent to begin the surgical dissection by encircling the ipsilateral main pulmonary artery and both pulmonary veins to obtain proximal and distal control in the event of vessel injury.

Because the pulmonary circulation is under low pressure, arterial and venous injury can usually be immediately controlled with local pressure at the injury site. It is necessary, when massive hemorrhage occurs, that the surgeon make an immediate analysis, knowing the site and magnitude of the injury, of what will be required to control the bleeding. This often requires calling for additional assistance in the operating room. With additional help and a good plan of attack, it is usually possible to readily obtain control, and perform a fine nonabsorbable suture. Rarely, injury to the main pulmonary artery, left atrium medial to the pulmonary vein, or superior or inferior vena cava will require cardiopulmonary bypass to control the situation for adequate repair.

In addition, injuries to bronchial arteries, parenchymal surfaces, pleural adhesions, as well as intercostal and internal mammary vessels can also lead to significant intraoperative blood loss if they go unnoticed. It is important to use careful dissection techniques at all times during the conduct of the operation to avoid injury to these structures and to control them quickly when they occur.

Ventilatory Complications A host of problems can occur and put the gas exchange of the patient at risk. If ventilation is established through a double-lumen endobronchial tube or a single-lumen tube with a bronchial blocking balloon, it is essential that the surgeon as well as the anesthesiologist be confident that it is in the correct position before starting the resection. The surgeon must also be aware of the presentation of tube displacement. High airway pressure and absent CO₂ in the ventilator circuit indicates that the bronchial cuff has herniated into the trachea producing an effective tracheal obstruction. Deflation of the cuff solves the problem and advancement of the tube prevents the problem from reoccurring. While conducting a right-sided resection with ventilation only on the left, persistent hypoxemia suggests that the left limb off the double-lumen tube has advanced too far and is occluding the left upper lobe orifice. This problem is sometimes first detected by the attentive surgeon, who recognizes that the usual ventilatory movement of the mediastinum is absent because of the progressively atelectatic left upper lobe. Withdrawal of the tube solves the problem.

Occasionally, it is necessary to change the endotracheal tube during resection. During resection of central T2 or T3 lesions, the endobronchial cuff may be injured or the tube may interfere with the bronchial resection or repair. The surgeon can assist with tube change by fixing a heavy suture to the distal end of the tube to be changed. After it is withdrawn, the anesthesiologist affixes the new tube to the suture. Traction on the suture by the surgeon helps guide the tube into position.

Patients undergoing thoracotomy for lung cancer resection as a rule are more susceptible to barotrauma pneumothorax because of preexisting bullous emphysema. Pneumothorax can occur at the time of induction and onset of positive pressure ventilation or, at any point, during the actual operation on the contralateral side. The surgeon should be aware of this development because airway pressures will increase, and the rhythmic movement of the mediastinum will be absent. Indeed, the mediastinum will sometimes balloon out toward the operative side. The problem is easily remedied by opening the mediastinal pleura.

Nerve Injury Nerve injuries occur in 1% to 2% of patients undergoing surgery for lung cancer. The recurrent laryngeal nerve as well as the phrenic nerve are generally considered the most important nerves encountered during lung cancer surgery. However, musculocutaneous nerves, such as the long thoracic nerve, can also be injured directly during dissection or by traction injuries during exposure. Injury to the phrenic nerve most often occurs when tumors are adherent or directly adjacent to the nerve with injury resulting in hemidiaphragm paralysis. Diaphragm plication may improve symptoms of dyspnea and breathlessness following phrenic nerve injury although the relative benefit remains controversial.²⁴

The recurrent laryngeal nerve provides ipsilateral motor innervation to the intrinsic laryngeal muscles for vocalization. Injury to the nerve unilaterally may result in hoarseness or a weak voice, whereas bilateral nerve injury frequently results in a compromised airway with associated respiratory distress in the immediate postoperative period. The right recurrent laryngeal nerve can be injured during resection of apical tumors on the right as it originates from the vagus nerve at the level of the right subclavian artery medially. The left recurrent laryngeal nerve can be injured during mediastinoscopy as well as during dissection of level 5 and 6 lymph nodes or during dissection of tumors adjacent to the preaortic and subaortic region on the left side. Unilateral nerve injury significantly increases the risk of aspiration and pneumonia following lung resection. Direct laryngoscopy is diagnostic, and strict precautions should be undertaken to avoid aspiration when the diagnosis is made. Several surgical techniques are available to improve vocal cord function after injury if symptoms are debilitating or if they do not improve with speech therapy.

General Considerations There are many complications that can arise after pulmonary resection, and some can be fatal if not recognized and managed early and aggressively. Attention to patient symptoms, clinical exam, and routine chest x-rays (CXR), in addition to a high level of suspicion during the postoperative period, will help the clinician identify complications and manage them effectively.