Protective lung ventilation during general anesthesia: is there any evidence?

In acute respiratory distress syndrome (ARDS) several studies have shown that mechanical ventilation with high tidal volume (VT) and low levels of positive end-expiratory pressure (PEEP) can promote ventilator-induced lung injury (VILI), thus increasing morbidity and mortality [1]. An open lung strategy, combining the use of low VT with adequate PEEP levels and recruitment maneuvers, has thus been recommended in ARDS patients [2–4]. In patients without ARDS admitted to intensive care units (ICUs), who required mechanical ventilation for at least 12 hours, the use of a high VT significantly increased the inflammatory response [5, 6]. In contrast to critically ill patients, during general anesthesia, mechanical ventilation is required only for a few hours, thus the beneficial effects of lung-protective ventilation remain questionable. Moreover, there are limited data from few randomized controlled trials with only small cohorts of enrolled patients.


Introduction
In acute respiratory distress syndrome (ARDS) several studies have shown that mechanical ventilation with high tidal volume (V T ) and low levels of positive endexpiratory pressure (PEEP) can promote ventilatorinduced lung injury (VILI), thus increasing morbidity and mortality [1]. An open lung strategy, combining the use of low V T with adequate PEEP levels and recruitment maneuvers, has thus been recommended in ARDS patients [2]- [4]. In patients without ARDS admitted to intensive care units (ICUs), who required mechanical ventilation for at least 12 hours, the use of a high V T signifi cantly increased the infl ammatory response [5], [6]. In contrast to critically ill patients, during general anesthesia, mechanical ventilation is required only for a few hours, thus the benefi cial eff ects of lung-protective ventilation remain questionable. Moreover, there are limited data from few randomized controlled trials with only small cohorts of enrolled patients.
Two recent meta-analyses that enrolled patients from ICUs and the operating room (OR) showed that lungprotective ventilation was associated with lower mortality and postoperative complications [2], [7]. However, there are no recommendations regarding optimal ventilatory strategies in patients without lung injury during general anesthesia.
In the present article, we provide a comprehensive picture of the current literature on lung-protective ventilation during general anesthesia in patients without ARDS, focusing on the applications of this strategy in patients undergoing abdominal, thoracic and cardiac surgery.

How mechanical ventilation is applied in the operating room
Although the protective ventilation approach may be benefi cial in a broader population with and without ARDS, the use of high V T without PEEP is still common during general anesthesia. A large French multicenter observational study, in which more than 2,900 patients undergoing general anesthesia were enrolled, showed that 18 % of patients were ventilated with a V T greater than 10 ml/kg body weight and 81 % without PEEP [8]. Moreover, a recruitment maneuver was applied in only 7 % of patients.
Similarly a 5-year observational study, in which 45,575 patients were enrolled, reported that although use of a V T less than 10 ml/kg and PEEP levels greater than 5 cmH 2 O increased progressively over time, 16-18 % of patients continued to receive a V T greater than 10 ml/kg without application of PEEP [9]. Th e presence of obesity and a short height were the main risk factors for receiving a large V T during prolonged anesthesia [10].

Rationale for lung-protective ventilation during general anesthesia
General anesthesia aff ects lung function primarily because of the loss of muscle tone, which promotes a reduction in lung volume, an alteration in ventilationperfusion ratio and the onset of lung atelectasis. Th e development of atelectasis is very common and occurs in more than 90 % of subjects undergoing general anesthesia [11], [12]. Atelectasis is mainly due to three basic mechanisms [13], [14]: • compression atelectasis • absorption atelectasis • loss of surfactant atelectasis.
Compression atelectasis is caused by the alterations in chest wall mechanics induced by general anesthesia per se and by several other mecha nisms, such as the patient's position (head-down), the body mass index, the age of patient and the type of surgery (abdominal surgery or laparoscopy), which increase intra-abdominal pressure (IAP), thus decreasing chest compliance and functional residual capacity (FRC), with the consequent development of intraoperative atelec tasis, intrapulmo nary shunting and hypoxemia. Other factors related to surgery can contribute to the reduction in pulmonary infl ation and to the development of atelectasis, such as a prolonged recumbent position intraoperatively, residual pain that reduces cough eff ective ness, and postoperative diaphragmatic dysfunc tion that can persist for up to one week [15], [16]. If the FRC is reduced below closing capacity, airway closure will occur; consequently the lung bases will be well perfused, but underventilated due to airway closure and alveolar collapse. Th is phenomenon increases ventila tion-perfusion mismatch and promotes further atelec tasis generation and hypoxemia.
Absorption atelectasis can be caused by exposure to high inspired fraction of oxygen (FiO 2 ) levels. When oxygen is absorbed from the alveolar gas into the capillary in distal occluded alveolar areas or where the ventilation-perfusion ratio is low or high FiO 2 levels are delivered, reabsorption of gas is promoted and generates atelectasis [11].
Loss of surfactant atelectasis arises from alterations in surfactant induced by eff ects of general anesthesia on healthy lungs [17].
Th e presence of atelectasis is an important factor in the pathogenesis of postoperative pulmonary complications, such as hypoxemia, pulmonary infections and local infl ammatory response [18]. Postoperative pulmonary complications in the fi rst hours after surgery are mainly due to atelectasis in the dependent regions of the lungs. Lung atelectasis may also promote the development of VILI by lung overdistension and by cyclic opening and closing of lung units at the boundary between the normally infl ated and collapsed lung units. On the basis of several studies of mechanical ventilation in ARDS patients, the same mechanisms of injury could be applied to mechanically ventilated patients during general anesthesia with healthy lungs. Th e use of recruitment maneuvers associated with adequate levels of PEEP could open and keep open previously collapsed lung regions. In addition, the use of a low-moderate V T could avoid overstress-overdistension of lung units.

Protective versus conventional lung ventilation strategies during general anesthesia
In Table 1, we provide a synopsis of randomized controlled trials (RCT) comparing protective versus conventional lung ventilation strategies during general anesthesia over time, in specifi c surgical settings showing the main outcomes explored in these studies. In Figure 1, we show the numbers of RCTs that we considered, divided according to the type of surgery.

Abdominal surgery
Postoperative pulmonary complications remain a significant problem after surgery. Th ey occur in 5-10 % of all surgical patients and 9-40 % of those undergoing abdominal surgery experience postoperative pulmonary complications [19], which increase morbidity and mortality [19], [20]. Among the postoperative pulmonary complications, lung atelectasis is one of the principle mechanisms for the development of VILI, pneumonia and postoperative respiratory failure.
In this context, Wrigge and colleagues investigated in two studies the eff ect of diff erent ventilatory strategies on the release of infl ammatory mediators in patients undergoing elective surgery [21], [22]. In the fi rst study, 39 patients scheduled for extra-thoracic surgery (abdominal, vascular, bone and other) were randomized to one of three mechanical ventilation strategies: 1) V T of 15 ml/kg ideal body weight without PEEP; 2) V T of 6 ml/ kg without PEEP; and 3) V T of 6 ml/kg with PEEP 10 cmH 2 O. Plasma levels of interleukin (IL)-6, IL-10 and tumor necrosis factor (TNF)-α were measured after one hour of mechanical ventilation [21]. In the second study, 64 patients undergoing general anesthesia were randomized to receive mechanical ventilation with a V T of 12-15 ml/kg ideal body weight without PEEP, or with V T of 6 ml/kg and PEEP levels of 10 cmH 2 O. Local and systemic infl ammatory biomarkers, including IL-8, IL-1, IL-6, IL-10, TNF-α and IL-12, were determined after 3 hours of mechanical ventilation [22]. Both studies were unable to fi nd any signifi cant diff erences in terms of infl ammatory mediators and the authors concluded that, in contrast to patients with acute lung injury in whom there is a systemic infl ammatory reaction during major surgery, in uninjured normal lungs short term mechanical ventilation alone with high V T levels did not increase pulmonary or systemic infl ammation related to surgery [21], [22]. No diff erences in biomarkers of lung epithelial injury were observed after 5 hours in a later study, which compared ventilation with V T 12 ml/kg ideal body weight without PEEP versus V T 6 ml/kg and PEEP 10 cmH 2 O [23].
To explore the eff ect of a high compared to a low V T for similar PEEP levels, Treschan et al. randomized patients to receive a V T of 12 ml/kg body weight versus 6 ml/kg with a PEEP of 5 cmH 2 O [24]. Except for the intraoperative oxygenation, which was higher in the high V T group, there was no signifi cant diff erence in forced vital capacity and forced expiratory volume in one second between groups, for up to fi ve days after the surgery.
Diff erent from the previous studies, Weingarten et al. evaluated an open lung strategy in which low V T ventilation was associated with PEEP plus a recruitment maneuver in order to minimize atelectasis and shear stress in the lung parenchyma [25]. Th is open lung strategy, consisting of a V T of 6 ml/kg predicted body weight with PEEP 12 cmH 2 O and recruitment maneuvers, signifi cantly improved only intraoperative oxygenation with no diff erence in the infl ammatory response or length of hospital stay compared to a V T of 10 ml/kg without PEEP [25]. Th ese fi rst studies seem to suggest that a protective ventilator strategy does not have any role in patients without lung injury [21]- [25]. However, these studies demonstrated that this mode of ventilation is feasible in open abdominal surgery with no adverse eff ects [23], [25]. In contrast to the previous studies, Severgnini et al., comparing a lung protective mechanical ventilation consisting of a V T of 7 ml/kg ideal body weight with PEEP levels of 10 cmH 2 O and recruitment maneuvers versus a V T of 9 ml/kg without PEEP, showed benefi cial eff ects of the lung-protective strategy during general anesthesia lasting more than 2 hours [26]. Th e lung-protective strategy improved postoperative respiratory function in terms of dynamic spirometry, oxygenation, and pulmonary complications for up to 5 days after surgery, without increasing the incidence of intraoperative complications. Although there was no signifi cant diff erence in the hospital length of stay between groups, 20 % of the patients in the lungprotective group, compared with 40 % in the control group, were still in hospital on postoperative day 14 [26]. A recent multicenter randomized clinical trial in which lung-protective ventilation with a V T of 6-8 ml/kg predicted body weight, PEEP 6-8 cmH 2 O and recruitment maneuvers repeated every 30 minutes was compared with non-protective ventilation with V T 10-12 ml/kg without PEEP, found that the lung-protective ventilation signifi cantly reduced major pulmonary and extrapulmonary complications from 27.5 % to 10.5 % [27]. Th e lung-protective strategy also signifi cantly reduced the proportion of patients who required postoperative ventilator assistance from 17 % to 5 % and the hospital length of stay.
Compared to the earlier studies [22]- [25], these two recent trials found a benefi cial eff ect of a lung-protective strategy probably because of the large number of enrolled patients, the homogeneity of the selected population of patients undergoing open abdominal surgery with an expected duration of at least 2 hours, the standardization of fl uid management, and the clinically relevant outcomes explored (not only lung infl ammatory mediators) in the postoperative period.
Th ese results demonstrate that in patients undergoing abdominal surgery a multifaceted open lung protective strategy can prevent the intraoperative alveolar opening and closing and overdistension of lung areas that lead to VILI and pulmonary complications. Currently, we are waiting for the results of the PROVHILO study, a worldwide multicenter RCT in which patients scheduled for abdominal surgery are being enrolled. In this study, all patients are ventilated with protective tidal volumes (in both groups, V T < 8 ml/kg predicted body weight) and randomly assigned to a lung-protective strategy with use of recruitment maneuvers and PEEP levels of 12 cmH 2 O or a conventional strategy without recruitment maneuvers and PEEP between 0 and 2 cmH 2 O [28]. If the results of this study confi rm those of the last two trials [27], [26], lung-protective strategies will be more widely applied in patients undergoing abdominal surgery [28].

Thoracic surgery
During thoracic surgery, one-lung ventilation is an established procedure that could increase the risk of promoting VILI compared to double lung ventilation, because of greater reduction in lung volume and greater degree of alveolar collapse in dependent lung regions. Two retrospective studies of patients who had undergone elective pneumonectomy found that larger intraoperative V T and higher inspiratory airway pressure were associated with the development of pulmonary edema and respiratory failure [29], [30]. Despite this, conventional mechanical ventilation in these patients consists of V T between 8-12 ml/kg to prevent lung atelectasis with zero or low levels of PEEP to avoid shunt aggravation by redistribution of blood fl ow to non-ventilated regions [31], [32]. However this approach is not an evidencebased guideline.  Schilling et al., in a randomized study in patients scheduled for open thoracic surgery undergoing onelung ventilation, showed that mechanical ventilation with V T of 5 ml/kg ideal body weight compared to 10 ml/kg signifi cantly decreased the pulmonary infl ammatory response up to 2 hours postoperatively [33]. Subsequently, Licker et al. retrospectively evaluated the implementation of a lung-protective ventilation strategy in lung cancer resection combining a low V T (< 8 ml/kg) with PEEP 4-10 cmH 2 O and recruitment maneuvers versus a conventional V T target ventilation of 9-12 ml/kg during two-lung ventilation and 8-10 ml/kg during onelung ventilation without recruitment maneuvers and PEEP applied at the discretion of the anesthetist [34]. Th e lung-protective strategy signifi cantly reduced the incidence of atelectasis (from 8.8 % to 5 %), postoperative acute lung injury (from 3.7 % to 0.9 %), ICU admission (from 9.4 % to 2.5 %) and length of hospital stay (from 14.5 ± 3.3 to 11.8 ± 4.1 days). Th ese data were confi rmed in a randomized study during elective lobectomy in which patients were ventilated with a high V T of 10 ml/kg without PEEP compared to a low V T of 6 ml/kg with 5 cmH 2 O of PEEP and pressure controlled ventilation [35]. Th e lung-protective ventilation was associated with a lower incidence of lung infi ltration or atelectasis (2 versus 10) and of cases of hypoxemia (1 versus 8).
During esophagectomy, a procedure requiring a prolonged period of one-lung ventilation, Michelet et al. demonstrated in an RCT that lung-protective ventilation (V T 9 ml/kg during two-lung ventilation, reduced to 5 ml/ kg during one-lung ventilation and PEEP 5 cmH 2 O throughout the operative time) could prevent alterations in lung function and reduce the infl ammatory response in patients without previous lung disease compared to conventional ventilation strategy (V T 9 ml/kg during twoand one-lung ventilation without PEEP) [36].
Th e majority of studies so far have demonstrated that, during thoracic surgery, traditional intraoperative ventilatory settings seem to be harmful. An intraoperative open lung approach based on small V T , moderate-high PEEP and recruitment maneuvers may be benefi cial but further randomized clinical trials are necessary to generate clinical evidence.

Cardiac surgery
In cardiac surgery, use of cardiopulmonary bypass (CPB), contact of the blood with artifi cial surfaces and ischemia/ reperfusion of the heart and lungs are associated with a pulmonary and systemic infl ammatory response, with activation of elements of the complement cascade, neutrophils and pro-infl ammatory cytokines [37]- [39].
Th is systemic infl ammatory response syndrome can be mild to severe in 10 to 35 % of cases and may induce an acute lung injury, which generally resolves within 24 hours. Th is clinical event contributes to increased morbidity and mortality [40]. In this context, injurious mechanical ventilation could aggravate the primary infl ammatory response described above (fi rst hit), representing a second hit. Moreover, during CPB, the lungs are not ventilated and either rest at low values of continuous positive pressure [41] or are completely disconnected from the ventilator [42]- [44]. Traditionally, ventilator settings in cardiac surgery patients included large V T (10-15 ml/kg) in order to minimize atelectasis and minimal levels of PEEP to reduce hemodynamic consequences. Following the results of clinical trials in ARDS patients [45], [46], there has been increased interest in protective lung ventilatory strategies during cardiac anesthesia and several trials have tried to demonstrate the role of protective lung ventilation in this context.
Koner et al. found no diff erences in plasma levels of IL-6 and TNF-α 2 hours after the end of CBP among patients randomized to receive protective ventilation (V T 6 ml/kg ideal body weight, PEEP 5 cmH 2 O) or conventional V T ventilation (V T 10 ml/kg) with and without PEEP levels at 5 cmH 2 O [42]. Th ere were no diff erences among groups in the explored clinical outcomes, including total intraoperative fl uid balance, intubation time and hospital length of stay [42].
Wrigge et al. measured pulmonary and plasma levels of diff erent cytokines and chemokines (IL-2, IL-4, IL-6, IL-8, IL-10, TNF-α and interferon-γ) in patients ventilated with high or with low tidal volumes (V T 12 ml/kg versus V T 6 ml/kg ideal body weight). Th ey observed higher values of TNF-α after 6 hours of ventilation with high V T , with no diff erences in other infl ammatory mediators [41].
However, a signifi cantly reduced infl ammatory response, in terms of pulmonary and systemic mediator levels (IL-6 and IL-8) was observed when applying a moderate PEEP level strategy (V T 8 ml/kg with PEEP 10 cmH 2 O) compared to a low PEEP and high V T strategy (10-12 ml/kg with PEEP 2-3 cmH 2 O) [44]. In comparison to previous studies [41], [42], this study evaluated a greater diff erence in PEEP levels and a longer duration of mechanical ventilation [44].
Reis et al. investigated the eff ect of open lung ventilation, consisting of low V T (4-6 ml/kg) with moderatehigh PEEP levels (10 cmH 2 O) and recruitment maneuvers, on infl ammatory mediators. In this study, they compared an early (immediately after intubation) and a late (at the end of CPB) application of the same open lung strategy, with conventional ventilation (V T 6-8 ml/kg, PEEP 5 cmH 2 O). Both the open lung approaches significantly decreased IL-8 and IL-10 levels after CPB [47]. Subsequently, the same authors showed that the early open lung approach signifi cantly attenuated the reduction in postoperative FRC, for up to 5 days after surgery, and reduced the incidence of hypoxemic events during the fi rst 3 days after extubation [48]. Ventilation and weaning times were similar among groups. Th is positive eff ect on postoperative FRC could be related to the prevention of additional lung injury caused by mechanical ventilation. Chaney et al. similarly reported better dynamic and static lung compliance and less shunt in patients ventilated with low compared to high V T (6 versus 12 ml/kg) [49].
Recently Sundar and colleagues observed that a larger number of patients were extubated after 8 hours (53 % versus 31 %) when ventilated with a low V T of 6 ml/kg ideal body weight compared to V T 10 ml/kg with similar PEEP levels. Furthermore, a lower postoperative reintubation rate was observed. However, global time to extubation was similar between groups, as were ICU length of stay and 28-day mortality [43].
Th ere is, therefore, a small amount of evidence from small studies in support of lung-protective ventilation in cardiac surgery patients [50]. However, the presence of several confounding factors, not related to mechanical ventilation, which could contribute to the development of a systemic infl ammatory response and postoperative pulmonary complications, may have infl uenced the main outcome results. Hence, further studies with larger cohorts of patients are needed to confi rm the still weak evidence in favor of lung-protective ventilation in cardiac anesthesia.

Conclusions
Mechanical ventilation is necessary for patients during general anesthesia. Although mechanical ventilation is considered a safe procedure, it can generate pulmonary stress and strain, promoting lung injury. Th ere is increasing evidence that lung-protective ventilation may be benefi cial in abdominal surgery (lower infl ammatory response and better outcome). During thoracic and cardiac surgery, lung protective ventilation has only been associated with a reduced infl ammatory response.
Lung-protective ventilation should be considered in the presence of pulmonary disease, prolonged anesthesia, in high-risk patients or for high-risk surgery. Although lung-protective ventilation may be benefi cial for the lung, it may impair the cardiovascular system, reducing venous return and cardiac output and requiring the use of fl uids and vasopressors. Th us, the risks and benefi ts of lungprotective ventilation need to be balanced in each individual patient.