This chapter will mainly review the ventilation strategies during thoracic surgery which when inappropriately applied may contribute to intraoperative lung injury. Although the inflammatory reaction of the lungs in response to ventilation or thoracic surgery may emerge and can be measured during one-lung ventilation, these do not predictably lead or correspond to the incidence of PPCs.^1,2 Therefore the relevant clinical hallmark of any intraoperative-inflicted lung injury is the increased incidence of PPC. Once present, PPCs contribute to an increased length of stay in high dependency units and may increase mortality and cost.^3,4

Characterizing Postoperative Pulmonary Complication

Because PPCs are used as a measure of intraoperative lung injury, we need to better define them before reviewing the literature. Postoperative morbidities which are defined in clinical trials as PPCs, consist of atelectasis, pneumonia, lung edema, poor oxygenation, need for supplemental oxygen, postoperative invasive or noninvasive ventilation, reintubation, pleural effusion, pneumothorax, acute lung injury, acute respiratory distress syndrome (ARDS), pulmonary aspiration, postoperative fever over 38°C, bronchospasm, pulmonary embolism, radiologic consolidation, and more.⁵ Usually, a composite subset of these signs has been used to increase the chances of identifying complications and help evaluate the severity of respiratory complications.^6,7 In one major study in a general population of surgical patients, Canet et al.⁶ used a carefully chosen set of such signs with appropriate definitions (Table 19.1), which was subsequently used in some studies to evaluate the incidence of postoperative lung injury after lung surgery.^8,9 There are some inherent problems in using the items of PPCs. Some of the items may have some degrees of overlap, that is, “pneumonia,” “consolidation,” and “atelectasis” or “respiratory failure” and “need for intubation.” Some of the complications, such as “need for oxygen” are considered a “minor” complication, whereas others such as acute lung injury or ARDS are “major” complications. Furthermore, the endpoints may or may not be appropriate depending on the clinical circumstances. For example, when evaluating the postoperative impact of intraoperative ventilation, endpoints such as atelectasis, pneumonia, need for oxygen, and pneumothorax may seem pathophysiologically linked to mechanical ventilation. Although aspiration pneumonitis or pulmonary embolism may contribute to PPCs as such, they may not necessarily be related to intraoperative ventilation. In general, linking specific intraoperative injurious interventions, such as mechanical ventilation, to PPCs may be difficult if they are not adequately controlled for unrelated intraoperative events (major bleeding, extent of tissue injury) or related to postoperative events (wound infections, anastomoses leakage, peritonitis), which may independently lead to PPCs. Thus consensual definitions of outcome measures in view of their pathophysiologic relevance have been attempted to standardize outcome measures for clinical studies.

At least two consensus papers on the issue of PPC and its use in perioperative studies have been published with quite different results. In 2015, a European Society of Anaesthesiology and European Society of Intensive Care Medicine (ESA/ESICM) task force published outcome measures for postoperative outcomes which included PPCs.⁷ The ESA/ESAICM closely followed Canets’ items and definitions of PPC issued in 2010,⁶ with minor changes (see Table 19.1). A more recent consensus statement by Abbott et al.¹⁰ used a systematic review and a three-stage Delphi process to choose appropriate items. Briefly, items were first procured from relevant available literature. Then, in three consecutive steps, investigators assessed and scored the items in terms of quality and validity, excluded items with low scores, and ultimately defined only four items: atelectasis, as detected with chest x-ray or chest computer tomography; pneumonia, using U.S. Centers for Disease Control and Prevention criteria; acute respiratory distress syndrome (ARDS) using the Berlin definition;¹¹ and pulmonary aspiration using “clear clinical history and radiologic evidence.” Unfortunately, none of the two consensual agreements are used consistently among investigators studying PPCs. Despite the controversies on items included in endpoint of PPCs among studies, it would seem prudent to at least relate some of the intraoperative issues under assessment, that is, intraoperative lung ventilation to pathophysiologically related and well-defined postoperative complications.

The Incidence of Postoperative Morbidity and Mortality in Thoracic Surgery

The incidence of PPC following lung resection or thoracic surgery varies in accordance with the data surveyed and items used to define the entity of the pulmonary complication. Licker et al.,¹² as well as Serpa Neto et al.¹³ used the definitions of acute lung injury and ARDS as a measure of PPCs and found an incidence of 4.2% and 4.3%, respectively in thoracic surgical patients. Significantly, once a patient suffered from these complications, the mortality rate rose to 26%. In a study using a relevant clinical database, Blank and colleagues¹⁴ used postoperative events such as tracheostomy, empyema requiring treatment, pneumonia, reintubation, initial ventilator support greater than 48 hours, ARDS, bronchopleural fistula, pulmonary embolism, air leak greater than 5 days, atelectasis requiring bronchoscopy, and respiratory failure to define their primary outcome of “respiratory complications” and found an incidence of 18% among patients undergoing pneumonectomy. Adjusting for patient characteristics, such as age, gender, body mass index, and American Society of Anesthesiologists status, the length of hospital stay in patients who developed respiratory failure and respiratory complications was increased by a factor of 4.7 and 3.5 respectively (Fig. 19.1). Alam et al.⁴ found an incidence of lung injury consisting of pneumonitis, acute lung injury, and ARDS of 5.3% after thoracic surgery; they report a 25% mortality rate in patients with PPC compared with 2.6% in those without. These studies show that postoperative pulmonary complications are highly relevant and, once developed, caused considerable increased in mortality.

Predisposing Condition for Postoperative Pulmonary Complication

Preoperative factors, such as age, preoperative peripheral oxygen saturation (SpO₂), respiratory infection before surgery, and preoperative anemia have been identified as risk factors contributing significantly to the development of PPCs.⁶ Some of these factors (smoking, preoperative lung function, preoperative infections) are potentially amenable to preoperative improvement with proper preoperative rehabilitation, which, if pragmatically practicable, would be mandatory before surgery (see Chapter 9).⁵

Intraoperative Lung Injury and Postoperative Pulmonary Complications

Intraoperative events which may injure or initiate injury in the lungs during the surgical procedure can be related to the surgery itself and/or to the anesthesiologic management. In the individual patient, both disciplines may contribute in a varying degree to PPCs. Some studies consistently show that both short- and long-term survival of lung cancer patients after surgery depend on the volume or number of operations performed at hospitals.¹⁵ Pezziet al.¹⁶ evaluated this issue in the National Cancer Database for major lung resections from 2007 to 2011. There were 124,418 major pulmonary resections identified in 1233 facilities. The 30-day and 90-day mortality rates were, respectively, 2.8% and 5.4%. Hospital volume of cases was significantly associated with 30-day mortality, with a ­mortality rate of 3.7% for an annual volume of cases less than 10, and 1.7% for a volume of 90 or more cases. The lowest volume centers had a 1.5- to 2-fold increase in 30-day and 90-day mortality in comparison to highest volume centers. This finding is significant considering that other improvements in care usually do not translate to a large difference in outcome.

Surgical Causes of Lung Injury

Surgery leads to a systemic inflammatory response, which involves the activation and interaction among cytokines and inflammatory cells, which contribute considerably to alveolar dysfunction.^17,18 The extent of surgery has been consistently shown to be closely associated with an increased inflammatory response^19,20 and increased perioperative morbidity.²¹ Salati M et al.²² report from the European Society of Thoracic Surgeons Database 2017 on patterns of care and perioperative outcomes of surgery for malignant lung neoplasm in 62,774 patients submitted to lung resections. They show that cardiopulmonary complications and mortality increase with the extent of the lung resection, respectively, 11.5% and 1.6% in segmentectomy, 17.6% and 2% in lobectomy, and 26.4% and 6.7% in pneumonectomy. The overall mortality in lung resection surgery is about 3%. Surgical approach is important, and although no large scale multicenter, randomized controlled clinical trials comparing minimally invasive video-assisted lobectomy (VATS) to open thoracotomy has so far been conducted, most retrospective studies using propensity matching show lower PPCs in the VATS group.²³ Lower systemic inflammatory response, lower pain scores, and shorter duration of chest tubes in the VATS group may be the cause of the positive short-term effects and a decrease in length of hospital stay. Blank et al.¹⁴ show that increased blood transfusion is an independent predictor of increase in PPC in patients undergoing pneumonectomy. In summary, the extent of surgery and surgical stress is a major contributary cause of PPCs after lung surgery. When there is a choice, decreasing the extent of surgery and surgical stress might lead to decrease in postoperative complications.

Anesthesiologic Causes of Lung Injury

In recent years, mechanical ventilation has been the focus of utmost attention in preventing intraoperative lung injury, which may then be manifested in PPCs. Mechanical ventilation as a possible cause of intraoperative lung injury will therefore be the focus of this chapter. Anesthetic regime, which may modify intraoperative lung inflammatory response and possibly postoperative morbidity. Other factors, especially intraoperative fluid administration (see Chapter 21) and postoperative sequelae of muscle relaxation are contributing factors.²⁴

Pathogenesis of Intraoperative Lung Injury Through Mechanical Ventilation

During mechanical ventilation, kinetic energy is necessarily expended on lung tissue to inflate the lung. However, kinetic energy may also overstretch and/or deform lung structures to potentially damage the lung. Stress and strain are concepts used to measure these changes.^25–27 Briefly, stress is defined as force per unit of area as induced by an external force applied onto a specific material. In terms of pulmonary physiology, the transpulmonary pressure (difference between alveolar and pleural pressure) would best represent stress. Strain is defined as the consequent change in dimension over the initial or resting value of that specific material. Strain during mechanical ventilation may be defined as the change in tidal volume relative to functional residual capacity (FRC). Stress and strain are conceptually related and are linked mathematically (stress = constant × strain).

How much strain (tidal volume in relation to FRC) is needed to injure a healthy lung? Because strain is related to tidal volume, Protti et al.²⁵ studied the effects of increasing levels of tidal volume corresponding to strain levels of 0.45 to 3.3 on lung injury in a large-animal study. They found that healthy pigs whose lungs were ventilated for over 50 hours with tidal volumes equivalent to strain values less than 1.5 (i.e., tidal volumes < 1.5 times the volume of the FRC in healthy pig lungs) showed neither physiologic nor histologic signs of lung injury. This is an interesting and intriguing finding because most recent studies on protective ventilation during surgery consider low tidal volumes in the approximate range of 0.20 to 0.25 of FRC (assuming a FRC of 2000–2200 mL) protective, and those in the range of 0.35 to 0.40 of FRC potentially injurious. Both ranges are far below the levels this study deems noninjurious in subjects with healthy lungs. However, atelectasis, which is prevalent during anesthesia, and anesthesia as such, will decrease the FRC, which might then make a low tidal volume a realistic option within a comprehensive protective ventilation strategy.

When is stress injurious? Stress is especially injurious when it is not evenly distributed throughout the healthy lung but concentrated in small areas.²⁸ Atelectasis, that is not recruited, leads to a cyclic concentration of stress in the interface between the aerated and nonaerated lung areas during inspiration and expiration, and is thus an important stress potentiator during mechanical ventilation.²⁸

Clinically, the injurious aspects of mechanical ventilation are also characterized by the terms barotrauma, volutrauma, and atelectrauma. Barotrauma is the injurious effects of high airway pressure leading to excess stress deforming and disrupting lungs tissue. Although barotrauma was previously used for any lung injury attributable to mechanical ventilation,²⁹ today it is reserved for direct effects of pressure on lung parenchyma leading to airway rupture, pneumothorax or pneumomediastinum. In pioneering studies,^30,31 Dreyfuss et al. showed that ventilating the lungs of rats with high airway pressure led to overt lung injury when the high airway pressure translated into high tidal volumes. The same level of high airway pressure alone did not lead to lung injury when it was deterred from leading to high tidal volumes by straps around the chest of the animals. The authors concluded that tidal volume and not airway pressure per se was the injurious factor and coined the term “volutrauma” for this entity.³² Further studies by other groups showed that low tidal volumes applied without positive end-expiratory pressure (PEEP) may promote atelectasis and allow repetitive recruitment and derecruitment of lungs, and thus lead to or potentiate lung injury.^33,34 Repetitive cyclic recruitment and derecruitment of atelectatic alveoli generates shear forces (in terms of stress), which lead to lung injury. This injurious mechanism of lung injury, theoretically prevalent during anesthesia, is called atelectrauma. Recently, the various damaging aspects of mechanical ventilation have been subsumed under one, all encompassing, “mechanical power.”^35,36 The higher the mechanical power expended on the lung, the higher the chance that the lung will be injured. The components of mechanical power are not only pressure (stress) and volume (strain) but also respiratory rate (stress and strain increase with rate) and the rate of inspiratory pressure change. This model attempts to provide a comprehensive explanation of the mechanical causes of lung injury and may in the future help us in better understanding of the potential risk when applying mechanical ventilation.³⁷

The damaging mechanical effects of ventilation lead to mechanical disruptions or breaks in alveolar endothelial and epithelial cell membranes^38,39 (Fig. 19.2). Any damage to these cells activates a local inflammatory response, which although ultimately needed in the repair process of these injuries, may first lead to release of cytokines, activation and recruitment of inflammatory cells, and release of enzymes and active oxygen species, which intensify the injury.⁴⁰ This inflammatory reaction in the lung is the primary substrate of intraoperative lung injury which may ultimately lead to PPC (see Chapter 18 for a more detailed account of this topic). It is therefore imperative to define and to implement a ventilation strategy which would avoid mechanical and inflammatory injury to the lung.

Mechanical Ventilation During Anesthesia and Lung Injury

Two pathophysiologic aspects of lung injury have been given substantial attention in defining new protective ventilation strategies during anesthesia and surgery.

The first is lung injury through atelectrauma, which develops when atelectasis has developed in the ventilated lung. Atelectasis, as supported by clinical and experimental evidence, has been long recognized as a very frequent and serious perioperative problem.^41,42 In the conscious subject, thoracic and diaphragm muscle tone and body posture are major determinants of the FRC, which is the lung volume at which the balance between the natural tendency of the lung to retract and that of the rib cage to expand occurs.⁴³ In the conscious subject during spontaneous ventilation, physiologic variability in tidal volume, variability in respiratory cycles,⁴⁴ and frequent sighs^45,46 superimposed on a lung kept inflated by chest wall and diaphragm muscle tone prevent atelectasis. During anesthesia, with or without paralysis, the chest and diaphragm muscle tone is decreased leading to lung compression and decreased FRC.⁴⁷ Compression of the lung leads to small airway closure and subsequent resorption of air and especially oxygen within the nonventilated alveoli, which will cause alveolar collapse and atelectasis. Lung compression and airway closure is a main determinant for the development of atelectasis and is highlighted by the fact that any further increase in lung compression increases atelectasis. Examples of increase in lung compression may occur in obese patients,^48,49 owing to the increased weight and pressure of thoracic and abdominal fat, and during laparoscopic abdominal operations⁵⁰ from increased lung compression through increased intraabdominal pressure. Beyond the more common atelectasis caused by compression of airways during anesthesia, atelectasis may also occur when an airway is blocked by tumor, pus, or mucus plug. Ventilation with high fraction of inspired oxygen (FiO₂) increases the incidence and prevalence of atelectasis because in alveoli distal to airway closure, oxygen more readily passes across a concentration gradient from the alveoli to the capillaries, whereas nitrogen, once it reaches a concentration equilibrium does not and will keep the alveoli open.^41,51

Atelectasis during intraoperative ventilation may cause lung injury by allowing repetitive cyclic opening and closing alveoli and thus generating shear trauma.^28,33,34 In addition, with parts of the lung now atelectatic, the remaining lung would overstretch with large tidal volumes.⁵² Finally, atelectasis formed during the intraoperative period may, in vulnerable surgical patients, persist beyond the intraoperative period with an increased incidence of PPC.^5,53

The second mechanism of intraoperative lung injury is high tidal volumes leading to alveolar overstretch and thus volutrauma. A significant event in the history and practice of mechanical ventilation was the publication of the ARDS Network study.⁵⁴ In that study, it was shown that low tidal ventilation (6 mL/kg predicted body weight [PBW]) as compared with 12 mL/kg PBW decreased mortality and increased ventilator-free days. In fact, the study was terminated prematurely, because considering the preliminary results, it was felt that it was unethical to ventilate the lungs of the patients at 12 mL/kg PBW. Anesthesiologists reasoned that because low tidal volumes are evidently beneficial for critically ill patients with severe lung injury, they may likewise be beneficial and should be applied in patients with healthy lungs undergoing surgery. However, the ARDS lung is not comparable with a healthy lung of a surgical patient. First, extensive consolidation of lung tissue in ARDS patients leaves only a small amount of lung able to accommodate the tidal volume. Because the amount of lung open to ventilation is as small as the size of a lung of a baby, the concept of a “baby lung” was introduced over 15 years ago by Gattinoni.⁵⁵ Using large or even moderate tidal volumes would excessively overinflate the baby lung (Fig. 19.3). Second, systemic and local lung inflammation present during ARDS would amplify any lung injury occurring in these open units.⁵⁶ Clearly, these pathophysiologic principles can not be translate to the mechanical ventilation in surgical patients with healthy lungs. Because most of the data addressing the effect of protective lung ventilation strategy are derived from studies conducted on critically ill, intensive care patients, there is no evidence to confirm that the same ventilation strategy is applicable to patients undergoing a general surgical or thoracic surgical procedure. Patients presenting for thoracic surgery, unlike patients suffering from ARDS, do not have areas of the lung that cannot be recruited and have no systemic inflammation. In fact, many thoracic surgery patients have a lung lesion that was discovered during a routine screening evaluation in an otherwise healthy individual.

The previously cited animal study by Protti et al.²⁵ on strain and tidal volume, as well as the reassessment of the ARDs trial by Amato et al.⁵⁷ on driving pressure (see later), clearly demonstrates the shortcoming of pinpointing tidal volume and the accompanying strain as the only culprit. High tidal volumes may lead to lung injury in a healthy lung as well when other conditions such as significant atelectasis, are present, and high tidal volumes may then potentiate atelectrauma.^28,58 The many studies applying low tidal volumes during anesthesia in patients with mostly healthy lungs, although not always conclusive, have over the past decade nevertheless led to a net decrease in tidal volume, but fortunately also in an overall change of ventilation practice,⁵⁹ which is a welcome improvement in quality of care.