99: Lung Volume Reduction Surgery

Lung volume reduction surgery (LVRS) is one of the most interesting and controversial areas in thoracic surgery. The purpose of the operation is to palliate dyspnea and improve functional status and quality of life for highly selected patients with emphysema. Chronic obstructive pulmonary disease (COPD) affects approximately 16 million Americans and is the fourth leading cause of death in the United States.¹ Worldwide there are estimated to be a billion smokers and, thanks to a global increase in the number of smokers each year, COPD is projected to be the third leading cause of death by the year 2020.² When pulmonary function tests demonstrate a forced expiratory volume in one second (FEV₁) of less than 30% of predicted values, a patient's 3-year mortality risk has been estimated at 40% to 50%. While medical therapy remains the mainstay of treatment for these patients,³ no medical therapy is able to improve pulmonary function or reverse the progressive nature of the disease. Three situations have emerged in which surgery is useful to palliate emphysema: lung transplantation, bullectomy, and LVRS. This chapter addresses the LVRS strategy.

The goal of LVRS is to palliate some of the distressing symptoms and limitations imposed by end-stage emphysema. Past controversy around this operation has focused on the procedure, interpretation of the results of trials and case series, issues about how new surgical procedures should be introduced and scientifically evaluated, and questions about how they should be funded by health care providers. Ideal candidates for LVRS have marked hyperinflation and significant regions of severe destruction with other distinct areas of more well-preserved lung parenchyma. The areas to be removed, frequently referred to as “target areas,” are usually, but not always, located in the upper lobes and have little pulmonary perfusion when studied with contrast CT or nuclear medicine perfusion scans. Surgical excision of these areas improves respiratory mechanics and function of the remaining lung. Clinically, the anticipated benefits are a reduction in dyspnea and improved exercise tolerance. A subset of highly selected patients may experience a survival benefit as well.⁴

Emphysema is characterized by abnormal permanent enlargement of air spaces distal to the terminal bronchiole accompanied by destruction of the airspace walls in the absence of obvious fibrosis.⁵ The destruction of pulmonary parenchyma causes a decreased mass of functioning lung tissue and also decreases the amount of gas exchange that can take place. As the lung tissue is destroyed, the lung loses elastic recoil and expands in volume. This leads to the typical hyperexpanded chest seen in emphysema patients with common findings including flattened diaphragms, widened intercostal spaces, and horizontal ribs. The increased distensibility of emphysematous lung results in a lung that is easily inflated but tends to remain pathologically inflated throughout the breathing cycle. An important consequence of this defect is that portions of severely emphysematous lung act as nonfunctional, volume-occupying areas. These anatomic changes result in the loss of mechanical advantages exploited in normal breathing and thus lead to an increased work of breathing and dyspnea.⁶ When the destruction and expansion occur in a nonuniform manner, the most affected lung tissue can expand disproportionally to compress and crowd the relatively spared lung tissue and thus impair ventilation of the functioning lung. Finally, there is obstruction in the small airways, likely caused by a combination of reversible bronchospasm and irreversible loss of elastic recoil by adjacent lung parenchyma. The suitability of a given patient for surgical treatment of emphysema depends in part on the relative contributions of lung destruction, lung compression, and small airways obstruction to the overall physiologic impairment of that patient. Sputum production, and the impact of excessive sputum on lung health and quality of life, would also play a role in predicting outcomes of LVRS.

The removal of severely diseased, poorly ventilated, and expanded lung tissue may decrease hyperinflation and improve the mechanical function of the diaphragm and thoracic cage. If the most severely diseased lung can be identified and resected, hyperinflation is reduced and overall breathing function will improve by restoring elastic recoil of the remaining lung. This will result in increased expiratory flow rates and allow the more normal lung to function without compression. LVRS may improve alveolar gas exchange and improve ventilation/perfusion mismatch. The end result is a marked improvement in respiratory mechanics and a decreased work of breathing, both of which help patients to function more normally and avoid the sensation of dyspnea. It is interesting to note that the true etiology or etiologies for dyspnea remain poorly understood. Although dyspnea is associated with severe airflow limitation, it is also linked to hyperinflation, respiratory muscle dysfunction, increases in respiratory drive, and abnormalities in alveolar gas exchange. Unfortunately, the correlation between the symptom of breathlessness and routinely measured physiologic parameters of pulmonary function are imprecise.⁷

The debilitating symptoms of pulmonary emphysema have attracted the interest of surgeons throughout the 20th century. Many creative but ineffective operations have been devised to treat the dyspnea caused by this disease. Costochondrectomy, phrenic crush, pneumoperitoneum, pleural abrasion, lung denervation, and thoracoplasty all have been offered up as surgical treatments for the hyperexpanded and poorly perfused emphysematous lung.⁸ As Laforet⁹ explained: “The alleged benefits of these maneuvers were frequently lost on patients whose worsening dyspnea left them with little energy to debate with their surgeons.”

LVRS was proposed by Brantigan¹⁰ in conjunction with lung denervation. Among 33 patients having the operation, there were 6 operative deaths (18% mortality) and no objective data to support the claim that survivors were subjectively better. LVRS was thus discarded after this initial experience showed the operation to be too risky with uncertain benefit. Over the following four decades, different groups attempted variations on Brantigan procedure with limited success. Observations about the physiologic behavior of emphysema patients during and after lung transplantation led to the reconsideration of volume reduction by Cooper.¹¹ Similar to Brantigan procedure, Cooper removed approximately 30% of the patient's lung volume by performing peripheral resection of the most emphysematous portions. However, the new approach used linear cutting/stapling devices and was performed as a bilateral procedure via median sternotomy. The procedure was designed to reduce dyspnea, increase exercise tolerance and performance in activities of daily living, and to improve quality of life.

Following this sentinel report by Cooper et al., LVRS enjoyed widespread application within the United States. A critical analysis of Medicare patients undergoing the procedure, however, revealed an unacceptably high mortality associated with the procedure, 23% at 12 months.¹² This led to cessation of Medicare funding for the operation and a decrease in enthusiasm for the procedure. As a way to rigorously evaluate the benefit of the operation, the National Institutes of Health sponsored a large, multicenter trial that began enrolling patients in 1999.¹³ The trial, called the National Emphysema Treatment Trial (NETT), was a prospective randomized study of 1218 patients and it has provided strong evidence for the efficacy, safety, and durability of LVRS.^4,14 The outcomes from the NETT analysis, in conjunction with data from earlier trials, help provide the criteria for defining which patients will benefit from LVRS.

History and Physical Examination

The evaluation of candidates for LVRS should be aimed at identifying patients with a physiologic profile that is most likely to respond to LVRS (Table 99-1). This includes patients with severe hyperinflation and reduced elastic recoil but less-pronounced airway disease, relatively well-preserved gas exchange, and no major comorbidities that carry an unacceptable perioperative risk.¹⁵ Because LVRS is a palliative procedure, the first evaluation step is to assess symptoms and the degree of quality of life impairment attributable to emphysema. The medical history needs to be thorough and include questions on daily activities and limitations caused by symptoms. The Medical Research Council (MRC) dyspnea score allows standardized grading of symptoms. Patients considered for LVRS typically have severe, incapacitating emphysema. They must have stopped smoking for at least 6 months and have received optimal medical management for their COPD, which includes pulmonary rehabilitation and nutritional support if necessary. Patients must be highly motivated to undergo surgical treatment and be willing to accept the risk associated with LVRS. The evaluation needs to differentiate COPD with predominant airways disease (chronic bronchitis and bronchiectasis) from COPD with predominant features of emphysema. Because the mean age of candidates is in the mid-60s, medical comorbidities are common and cardiovascular disease, cerebrovascular disease, obesity, and cachexia deserve particular attention. Patients with cardiovascular disease may be difficult to evaluate by elicited symptoms or signs as their emphysema limits their ability to induce anginal symptoms. It is important to note that exercise testing is often not useful because of the patient's inability to exercise to heart rate limits. Echocardiography may be limited by chest hyperinflation resulting in poor visualization of the heart. The use of dipyridamole or adenosine during cardiac testing is limited due to concern for inducing bronchoconstriction. Ultimately, because of the barriers to noninvasive evaluation, many candidates for LVRS undergo cardiac catheterization to obtain a definitive answer regarding their cardiovascular risk.

Pulmonary Function

The cornerstone of pulmonary function testing is spirometry. It is used to quantify the degree of airflow obstruction as well as its reversibility with bronchodilator drugs. Airflow obstruction is the most significant abnormality with emphysema and it can be estimated accurately by forced expiratory maneuvers. Lung volumes are measured by plethysmography rather than by dilution techniques because the latter measurements tend to underestimate the degree of trapped gas and residual volume. Additional parameters of pulmonary function that are assessed include resting and formal exercise arterial blood gas analysis. These are indicative of the patient's pulmonary reserve and reflect their potential for recovery after surgery. In addition, diffusing capacity, as measured by D_LCO values, estimates the severity of damage to the pulmonary capillary bed.

Exercise Capacity

The distance walked in 6 minutes is frequently used to assess exercise capacity. This test evaluates cardiopulmonary function and documents the amount of supplemental oxygen necessary to maintain oxygen saturation above 90%. It is also a useful parameter for the objective documentation of functional improvement following LVRS. Other approaches, including that used by the NETT, have used formal cardiopulmonary testing which results in a maximal exercise capacity expressed in Watts to provide an objective measure of functional capacity.

Radiologic Evaluation

The purpose of imaging in patient selection is to identify findings favorable for LVRS. The presence of hyperinflation, the severity of emphysema, and the distribution of emphysema are the main features to assess. These features are important based on the rationale for LVRS as well as to understand the relationship of preoperative radiographic findings with postoperative outcomes.¹⁶

The presence of hyperinflation is accurately assessed from inspiratory posteroanterior and lateral chest radiographs (Fig. 99-1A,B). Paired inspiratory and expiratory views show the maximum achievable diaphragm excursion and subjectively estimate the potential for improvement in diaphragm function if the lung volume is reduced by LVRS. The main indicators of hyperinflation are a low, flat diaphragm, increased AP diameter, and increased width of the intercostal spaces. Hyperinflated upper lobes may expand anterior to the upper mediastinum and increase the retrosternal space. Hyperinflation does not occur until emphysema is moderately severe and is a relatively specific sign for emphysema. Although hyperinflation should be present in a patient being considered for LVRS, there is no convincing evidence that the degree of hyperinflation predicts the surgical outcome.

The severity and distribution of emphysema on imaging studies correlates with clinical outcome after surgery. The best outcomes, as demonstrated by improvements in FEV₁ and exercise capacity, tend to occur in patients with more severe, heterogeneous disease that predominates in the upper lobes.^17,18 The most severe extreme of heterogeneous target areas is a giant bulla, but the term LVRS refers to lesser degrees of focal destruction in which some lung tissue is present that needs to be excised. The standard chest computed tomography (CT) examination is the most accurate means of evaluating the severity and distribution of emphysema (Fig. 99-2A,B). Unfortunately, there is considerable variation in the interpretation of CT scans in these patients, since the entire lung is affected to some degree by emphysema, and it may be difficult to assess the heterogeneity of the disease. Studies have demonstrated considerable interobserver variability in interpretation of the distribution of emphysema on CT scan.¹⁹ Despite this, the CT scan remains the central component of preoperative imaging. It is the most useful tool available for identification of potential target areas for resection, and also serves as a more rigorous screen than the radiograph for other underlying pathologies that would preclude LVRS. Lung cancer has been discovered in 2% to 8% of candidates for LVRS and some patients are candidates for combined cancer resection and volume reduction.^20,21 Pleural thickening and calcification raise concern for adhesions. In addition, bronchiectasis, inflammatory disease or asymptomatic infiltrates, and pulmonary hypertension may also be identified or suggested by CT and require additional investigation.

Nuclear medicine ventilation-perfusion lung scans depicting regional blood flow patterns provide a valuable roadmap for surgery. Although the absolute severity of emphysema is not accurately assessed, the presence of diffuse or upper or lower lobe predominant disease can be identified. Importantly, a right- or left-sided predominance of lung function may direct surgery toward a unilateral approach if corroborated by CT.

Review of surgical results of LVRS has identified risk factors for surgical mortality (Table 99-2). These predictors appear to reflect the function of different parts of the respiratory system. A markedly reduced diffusing capacity reflects a poor distribution of ventilation surface area to the available perfusion of the lung. The greater degrees of severity of disease are usually associated with a significant supplemental oxygen requirement and may indicate that the lung is too impaired to support the patient postoperatively. An elevated PaCO₂ indicates excessive work of breathing and chronic muscle fatigue and is a significant risk factor for complications after general anesthesia for any reason. It is not clear why upper lobe predominance has a better outcome than lower lobe resection. This is independent of the etiology for emphysema (smoking vs. alpha₁ – antitrypsin deficiency) and may be explained by the fact that the lower lobes comprise a larger volume of lung parenchyma compared to the upper lobes. Other markers of increased risk or poorer outcome include male gender, increasing age, and markedly reduced lung function with FEV₁ less than 20% of predicted values.

Historically, LVRS was accomplished via a median sternotomy. This is an approach that provides excellent exposure and a minimum of incisional morbidity. However, as comfort with thoracoscopic approaches to disease has increased, the majority of these procedures are now done via a VATS approach. We describe here a technique for performing a bilateral VATS in the supine patient, using buttressed staple lines. It is worth noting, however, that fairly uniform results have been obtained using a wide array of surgical strategies, including bilateral and unilateral approaches, open and thoracoscopic operations, and buttressed or unbuttressed staplers.

Many patients with severe emphysema have a significant element of chronic bronchitis with increased sputum production. Following induction of anesthesia, a single lumen tube is placed and flexible bronchoscopy is carried out to suction secretions and obtain a specimen for culture and for rapid gram stain to guide perioperative antibiotics. If thick, tenacious secretions are encountered, a minitracheostomy may be inserted at the end of the operative procedure to facilitate postoperative pulmonary toilet. Following bronchoscopy, the single lumen endotracheal tube is replaced with a left-sided double lumen tube.

The patient is positioned on the table with a padded roll beneath the shoulders, hips, and along the spine. The arms are then suspended, padded carefully above the patient's head, and affixed to a support bar. This position provides sufficient access to allow placement of three thoracoscopic ports; a camera port in the midaxillary line at the ninth interspace, an assistant port in the posterior axillary line at the eighth interspace, and a working port just medial to the anterior axillary line in the fifth interspace (Fig. 99-3).

Most candidates for LVRS have upper lobe predominant disease. Several minutes after ventilation is suspended to the right lung, the right middle and lower lobes are usually well deflated and become progressively atelectatic. The pulmonary ligament is divided and adhesions are taken down under direct vision. Dense adhesions are not common but may be encountered if there have been prior episodes of pneumonia. We, very rarely, might use an extrapleural dissection adjacent to adhesions to avoid injuring the fragile lung parenchyma.

For upper lobe disease, 70% to 80% of the right upper lobe is excised with multiple applications of a linear stapler buttressed with strips of bovine pericardium. It is often easier to apply the stapler to the deflated lung and this can rapidly be accomplished by using the cautery to fenestrate the apex of the right upper lobe (Fig. 99-4). The marked collateral ventilation leads to prompt collapse. A long, straight intestinal clamp can be applied to the lung to create a linear “crush” mark before application of the linear stapler (Fig. 99-5). We currently staple straight across the upper lobe beginning medially above the hilum and ending up just above the upper extent of the oblique fissure. Care should be taken to avoid crossing the fissure as this may damage the superior segment of the lower lobe (Fig. 99-6). In addition, stapling across the fissure may tether the superior segment of the lower lobe to the remaining upper lobe and prevent the superior segment from fully expanding and filling the apex of the chest. It is important to remember that the goal is to adequately reduce volume, not to remove all of the diseased lung.

Occasionally, the apex of the upper lobe will be densely adherent to the apex of the chest and to the superior mediastinum. In such cases, it may be easier to first transect the upper lobe as described above before attempting to dissect the apical and mediastinal adhesions (Fig. 99-7). Once the transection has been accomplished, the specimen can be more easily detached from the chest wall and mediastinum using blunt or sharp dissection, cautery, or even a linear stapler leaving a small remnant of the lung attached to the mediastinum if necessary. It is imperative to avoid injury to the phrenic nerves—a complication that will severely compromise postoperative recovery.

Following the upper lobe resection, the chest is partially filled with warm saline and the lung is gently inflated. Air leaks at this time are unusual but the reexpanded, remaining lung often does not completely fill the apex of the chest. We have explored the use of a pleural tent in this situation but now reserve it for rare instances, in particular when the remaining lung remains tethered in the chest by virtue of adhesions to the chest wall or diaphragm. It has not been our practice to perform mechanical or talc pleurodesis, even if the patient is not a potential candidate for subsequent lung transplantation. The efficacy of pleurodesis is questionable and the morbidity can be burdensome.

Two chest tubes are placed in the pleural space via the anterior and middle axillary port sites (Fig. 99-8). The posterior tube is brought across the dome of the diaphragm and halfway up the posterior chest. The anterior tube is brought to the apex of the chest near the mediastinum. Recently, the authors have been using a single chest tube after a straightforward resection with no intraoperative air leaks identified.

Ventilation is shifted from the left lung to the right lung and similar port sites are placed on the left side. With upper lobe predominant disease, the goal is to excise the superior subdivision of the left upper lobe leaving the lingula intact, as this is usually much less diseased. The left pulmonary ligament is usually divided but this requires displacement of the heart. If exposure is limited, the ligament is left intact and adhesions between the left lower lobe and the diaphragm are taken down. Unlike the anatomic situation on the right side, the superior segment of the left lower lobe usually reaches easily to the apex of the chest even without division of the pulmonary ligament.

The upper half to two-thirds of the left upper lobe is excised with multiple applications of the linear stapler. This is facilitated by deflating the apex of the upper lobe with cautery puncture. The long, straight intestinal clamp is often useful in helping to identify and demarcate the proposed line of excision. The line of excision is nearly parallel to the oblique fissure separating the upper and lower lobes. Similar to the right side, care is taken to avoid stapling across the fissure into the superior segment of the left upper lobe (Fig. 99-9).

Following left upper lobe excision, the lung is reinflated and inspected for air leaks. Two chest tubes are placed in a similar manner to the right-sided tubes. Virtually all patients are extubated in the operating room. If excessive secretions are present following the procedure or during the initial bronchoscopy, a 4-mm diameter minitracheostomy is inserted at the time of extubation to facilitate aggressive pulmonary toilet.

Pain control is essential to ensure adequate postoperative respiratory function. An epidural catheter is placed and it is imperative to confirm that a bilateral block has been established before beginning the procedure. Although this is not absolutely necessary, we use fluoroscopy to allow placement of the catheter at the tip of T4. During induction, hypotension can result from air trapping as well as pneumothorax. Air trapping leading to “pulmonary tamponade” is treated by removing the patient from the ventilator and allowing them to exhale to the atmosphere. Bronchodilators and prolonged exhalation times are also helpful but it is often necessary to remove the patient from the ventilator during exhalation.

During the procedure, our anesthesiologists avoid narcotics and limit the use of benzodiazepines to limit long-term respiratory depression. The inspiratory pressures are limited to 15 to 20 cm of water during one lung ventilation to avoid excessive pressure on the fresh staple lines. Most of these patients have an element of CO₂ retention and permissive hypercapnia is tolerated. During emergence, patients may be drowsy and it is not unusual to have the initial PaCO₂ as high as 90 mm Hg. This is transient and will come down as the patient receives nebulizers and chest physical therapy in the recovery room. It is important not to overreact and intubate, thus risking barotrauma, air leaks, and an unnecessary period of mechanical ventilation.

Contrary to the management of most patients following pulmonary resection, the chest tubes following LVRS are immediately placed to water seal without the application of suction. The loss of elastic recoil, the obstructive physiology of the remaining lung, and the fragile nature of the tissue makes the lungs more susceptible to hyperinflation caused by chest tube suction. The keys to postoperative management are adequate pain relief, aggressive chest physiotherapy, early ambulation, and management of secretions. Inhaled bronchodilators and systemic steroids during the perioperative period can be very beneficial in reducing airway reactivity.

The value of LVRS as a palliative procedure is clearly dependent on the surgeon's ability to minimize the frequency and severity of postoperative complications.²² The true mortality is difficult to assess due to bias in publishing case series with favorable results. Furthermore, individual case series often have an unintentional positive bias as patients that experience good outcomes are more likely to return for follow-up. Despite these caveats, it is noted that most of the published reports on the outcomes of LVRS describe a perioperative mortality under 8%.

The key to reducing morbidity is applying strict criteria at the time of evaluation. The risk factors for poor outcomes have been summarized in Table 99-2 and adhering to these criteria makes the selection fraction (patients selected for surgery divided by total patients evaluated) quite small. In the prospective trial reported by Pompeo,²³ only 60 patients were randomized to surgery or medical treatment, from the 237 patients that were initially evaluated. Similarly, in Ciccone's²⁴ report out of Washington University, nearly 3000 patients were evaluated to select the first 250 bilateral LVRS candidates. Fewer than one-third of patients referred for LVRS will be invited for onsite evaluation and only a fraction of these will be offered surgery. The NETT²⁵ demonstrated that performing LVRS on two high-risk subgroups: patients with a very low FEV₁ and homogeneous emphysema and patients with a very low FEV₁ and a low D_LCO resulted in an unacceptably high mortality. These patients have been viewed by many to be of too high risk for LVRS and these findings highlight the importance of prudent patient selection to maximize the benefit of this palliative procedure.

The most common complication following LVRS is a parenchyma air leak. Most patients have an initial air leak and it prolongs hospitalization in up to 15% of patients. Prevention strategies focus on using buttressed staple lines with pericardium or PTFE strips and ensuring pleural symphysis with a pleural tent when necessary.²⁶ In the absence of a large air leak, most patients will improve with time. Although chest tubes following LVRS are initially placed to water seal, suction is used for larger and symptomatic air leaks. If the air leak continues and prevents hospital discharge, the chest tube can be shortened and connected to a Heimlich valve as long as the remaining lung remains expanded off suction. Pleurodesis and reoperation are rarely indicated but are occasionally performed for persistent large air leaks to revise the staple line or excise the leaking site.

Infections are the second major class of complications following LVRS. Pneumonia is the most serious of these and is avoided by aggressive management of secretions and early ambulation. Sputum is routinely submitted for culture at the time of surgery so objective data is available if infiltrates progress postoperatively. Empyema is rare but may be severe as patients can have an air leak and a pleural space problem. However, low-grade pleural contamination may help to promote pleural fusion.

Respiratory decompensation and mucus plugging are a major problem following LVRS. Prevention measures focus on early ambulation and good respiratory therapy. High-risk patients with thick copious sputum are carefully scrutinized during their evaluation. If history and preoperative bronchoscopy are concerning in this regard, a minitracheostomy is placed following extubation in the operating room. If preoperative bronchoscopy reveals heavy, purulent sputum then the procedure should be canceled with a plan to reschedule after a course of antibiotics and reassessment.

Gastrointestinal complications are notable in patients following LVRS.²⁷ The link between emphysema and gastrointestinal complications is uncertain but many reports have noted a higher rate of intestinal ischemia, perforation, and ileus. We favor early colonoscopy to treat pseudo-obstruction and rule out mucosal ischemia. Despite the fact that cigarette smoking predisposes to both emphysema and cardiovascular disease, perioperative cardiac complications are relatively rare. The reason for this is the careful preoperative screening for coronary artery disease and pulmonary hypertension. Our routine is to perform a radionuclide ventriculogram on all patients and follow-up abnormal results with cardiac catheterization.

Many groups have reported preliminary results for LVRS, and these results have consistently shown benefit to the recipient with acceptable mortality and varying morbidity.^28–33 The consistent theme among reports of successful lung volume reduction programs has been meticulous patient selection, methodical patient preparation with reduction of risk factors, and attentive postoperative care. Most groups have reported operating on patients with a mean age of 65 years and a preoperative FEV₁ of 600 to 800 mL. The typical postoperative length of stay is 8 to 14 days with almost half the patients being detained due to persistent air leaks. Mortality ranges from 0% to 7% for the initial hospitalization. The expected benefits of the operation vary according to whether a unilateral or bilateral approach has been utilized, but gains of 20% to 35% in the FEV₁ have been reported for unilateral operations and gains of 40% to 80% are seen with bilateral operations. Most authors also report substantial gains in exercise tolerance, freedom from oxygen use, freedom from steroid use, and subjective quality of life.

As an example of the results reported from an individual institution, Ciccone²⁴ reported long-term results in 250 bilateral LVRS recipients. After a median follow-up of 4.4 years, the 5-year survival was estimated to be 68%. Eighteen of the 250 patients in that report proceeded to lung transplantation after a median interval of 4.3 years. Five years after surgery, the mean FEV₁ was 7% higher than that prior to surgery and half of the individual patients continued to demonstrate improvement over their preoperative FEV₁ values.

In 2003, the NETT reported the main results of a 5-year effort. This trial included 1218 patients randomized between LVRS and medical therapy between January 1998 and July 2002.⁴ In the entire cohort of patients, there was no difference in overall mortality between the surgical group and the medical therapy group. The surgical group did show significant improvement in exercise capacity, 6-minute walk distance, FEV₁, and quality of life measures. In subgroup analysis, a survival benefit was seen in the surgical arm for upper lobe predominant emphysema patients with low baseline exercise capacity, while a survival benefit was seen in the medical arm for nonupper lobe predominant emphysema patients with high baseline exercise capacity (Fig. 99-10). In long-term follow-up from this trial, published in 2006,⁴ the survival benefit in the upper lobe predominant emphysema patients with low baseline exercise capacity remained, but the survival difference between the medical and surgical arms in the nonupper lobe, high baseline exercise capacity group disappeared. Importantly, both the quality of life and exercise capacity improvements seen in the surgical group remained at long-term follow-up.

Early in the NETT experience, it was noted that patients with a FEV₁ less that 20% of predicted and either homogeneous distribution of emphysema on CT or a D_LCO of less than 20% of predicted had no change in exercise tolerance, no improvement in FEV₁, no subjective improvement in quality of life, and a 16% 30-day mortality after LVRS.²⁵ The high mortality found in this patient cohort prompted a modification of the NETT protocol, excluding from randomization any patients who met these criteria. However, retrospective review of a patient population where patients with homogeneous emphysema were excluded from consideration of surgery, but still met the high-risk criteria of the NETT protocol, demonstrated improved respiratory function, and an acceptable risk of mortality.³⁴ This suggests that the presence of suitable anatomic heterogeneity of disease may be the most important determinant of outcome. Indeed, the most common reason for exclusion of a patient from consideration of LVRS is the lack of sufficient target areas for resection.²⁴ For these patients, lung transplantation is the only surgical option that remains.

A commonly misunderstood result of the NETT is illustrated in Fig. 99-10D; namely, those patients with apical predominant emphysema and low exercise tolerance do better with surgery than medical therapy alone. Whereas there is a highly significant difference between these groups, the reason for the difference is not an improvement in surgical survival (e.g., compare Figs. 99-10D and E), but rather the high probability of death in the medical therapy arm. Until the NETT, the mortality of emphysema patients with low exercise tolerance was generally underestimated. This result also highlights the important role of control groups in evaluating new therapies.