Pulmonary Atelectasis: A Pathogenic Perioperative Entity

Three sets of mechanisms have been proposed that may cause or contribute to the development of atelectasis,4including compression of lung tissue, absorption of alveolar air, and impairment of surfactant function. This section describes these three underlying “physiologic” causes of atelectasis; clinical factors that can modulate the development of atelectasis are described in a subsequent section.

Compression atelectasis occurs when the transmural pressure distending the alveolus is reduced to a level that allows the alveolus to collapse. The diaphragm normally separates the intrathoracic and abdominal cavities and, when stimulated, permits differential pressures in the abdomen and chest. After induction of anesthesia, the diaphragm is relaxed and displaced cephalad and is therefore less effective in maintaining distinct pressures in the two cavities. Specifically, the pleural pressure increases to the greatest extent in the dependent lung regions (fig. 1) and can compress the adjacent lung tissue. This is termed compression atelectasis .5 

Several lines of evidence support a role of the diaphragm in this setting. Froese and Bryan6used cineradiography to demonstrate a cephalad shift of the diaphragm during anesthesia and spontaneous breathing, which did not progress after muscle relaxation. However, a difference in the pattern of diaphragmatic movement was noted. In supine patients, during spontaneous breathing, the lower, dependent portion of the diaphragm moved the most, whereas with muscle paralysis, the upper, nondependent part showed the largest displacement. Two distinctly different patterns of diaphragmatic displacement were seen from the same new functional residual capacity (FRC) position. In an anesthetized patient breathing spontaneously, the active tension in the diaphragm is capable of overcoming the weight of the abdominal contents, and the diaphragm moves the most in the lower, dependent portion (because the lower or posterior diaphragm is stretched higher into the chest, it has the smallest radius of curvature and therefore contracts most effectively). In addition, the diaphragm is thicker posteriorly than anteriorly, and this may account for the disproportionate movement.7During paralysis and positive-pressure ventilation, the passive diaphragm is displaced by the positive pressure preferentially in the upper, nondependent portion (where there is least impedance to diaphragmatic movement). Subsequent studies confirmed these findings and, in addition, documented a reduction in the transverse area of the chest.8Using an advanced computed tomography (CT) scanner, Krayer et al.  9demonstrated a reduced thoracic cross-sectional area in anesthetized subjects but had more variable results regarding shape and position of the diaphragm; some subjects showed a cranial shift of the diaphragm, but in other subjects, part of the diaphragm was unaffected or even moved caudally. Other investigators have also shown results inconsistent with the classic model of regional ventilation10; nevertheless, it can be concluded that FRC is reduced in the anesthetized subject, whether caused by loss of traction of the chest wall or compression of the lung. Loss of intercostal muscle function may also contribute to reduced FRC during anesthesia. Inhalation agents decrease intercostal muscle activity, particularly in children.11 

Hedenstierna et al.  8also noted an additional source of lung compression in that there was a net shift of central blood volume from the thorax, which seemed to pool in the abdomen, resulting in additional dependent pressure arising from the abdomen and acting on the diaphragm. Finally, the displacement of the diaphragm has been studied under dynamic conditions, whereby increases in diaphragm tension through phrenic nerve stimulation have been shown to reduce the amount of atelectasis at isovolumic conditions in anesthetized patients.12 

Therefore, compression atelectasis occurs during general anesthesia and is caused by chest geometry, overall cephalad diaphragm displacement, differential regional diaphragmatic changes, shift of thoracic central vascular blood into the abdomen, and altered diaphragmatic dynamics.

Resorption atelectasis—sometimes called gas atelectasis  4—can occur by two mechanisms. After complete airway occlusion, a pocket of trapped gas is created in the lung unit distal to the obstruction. Because gas uptake by the blood continues and gas inflow is prevented by blocked airways, the gas pocket collapses.13Under these conditions, the rate of absorption of gas from an unventilated lung area increases with elevation of the fraction of inspired oxygen (Fio²).14 

A somewhat different mechanism explains absorption atelectasis in the absence of airway occlusion. In this context, lung zones that have low ventilation relative to perfusion (low ventilation/perfusion [V^A/Q] ratio) have a low partial pressure of alveolar oxygen (Pao²) when air is breathed. When the Fio²is increased, Pao²increases, causing the rate at which oxygen moves from the alveolar gas to the capillary blood to increase greatly. The oxygen flux may increase so much that the net flow of gas into the blood exceeds the inspired flow of gas, and the lung unit becomes progressively smaller. Collapse is most likely to occur when the Fio²(and duration of exposure) is high or where the V^A/Q ratio (and mixed venous oxygen content) is low.15,16 

Pulmonary surfactant that covers the large alveolar surface is composed of phospholipids (mostly phosphatidylcholine), neutral lipids, and surfactant-specific apoproteins (termed surfactant proteins A , B , C , and D ). By reducing alveolar surface tension, pulmonary surfactant stabilizes the alveoli and prevents alveolar collapse. This stabilizing function of surfactant may be depressed by anesthesia, and such an effect has been confirmed in vitro  by Woo et al.  17The authors evaluated the effect of anesthetic agents on surfactant function using deflation pressure–volume curves in excised dog lungs. They found that the reduction in percent maximum lung volume was proportional to the concentration of both chloroform and halothane.17Wollmer et al.  18also used pulmonary clearance of technetium-labeled diethylenetriamine pentaacetic acid to demonstrate that halothane anesthesia, in combination with high oxygen concentration, caused increased permeability of the alveolar–capillary barrier in rabbit lungs. The authors postulated that the increased rate of pulmonary clearance of technetium-labeled diethylenetriamine pentaacetic acid during anesthesia with halothane was likely to be caused by combined effects on the pulmonary surfactant or the alveolar epithelium or both.18In addition, it is known that the content of alveolar surfactant in isolated lungs is modified by mechanical factors. Hyperventilation by increased tidal volume,19sequential air inflations to total lung capacity,20or even a single cycle of increased tidal volume19all cause release of surfactant in isolated animal lungs. In rabbits, maintained increases in tidal volume increased the amount of total phospholipids recovered from bronchoalveolar lavages.21,22Supporting this is the report that the spontaneous occurrence of large gasping respirations increases the proportion of active forms of alveolar surfactant (phospholipids). Oyarzun et al.  23examined the ventilatory variables of cats breathing spontaneously during anesthesia for 4 h. They found that the frequency of large gasps is directly correlated with the concentration of phospholipids in bronchoalveolar lavage fluid.23 

All three mechanisms—compression, gas resorption, and surfactant impairment—may contribute to atelectasis formation during general anesthesia (fig. 2). However, given the surfactant reserve and the 14-h surfactant turnover time, it may be that primary changes in surface forces are less important; it is not known whether a collapsed alveolus can denature surfactant, and so the 14-h turnover time may not be relevant. Nonetheless, absorption and compression are considered to be the two mechanisms most implicated in perioperative atelectasis formation.24 

The following section reviews the characteristics of atelectasis in several clinical contexts and then reviews contemporary concepts regarding the nature of atelectasis.

Anesthesia-induced atelectasis is traditionally thought of as collapse of alveoli. Brismar et al.  5showed that within 5 min after induction of anesthesia, areas of increased density appeared in the dependent regions of both lungs on CT. The dense areas had an attenuation factor that corresponds to blood and connective tissue and indicates the absence of air. Injection of radiocontrast in the pleural space showed that the densities were located above the pleura, i.e. , within the lung.25Hedenstierna et al.  26also found these densities in anesthetized sheep and confirmed histologically that the densities were collapsed lung regions and not fluid accumulation.

Methods to restore normal FRC and various reexpansion maneuvers have been suggested. The application of positive end-expiratory pressure (PEEP) has been tested in several studies. Arterial oxygenation does not usually improve markedly, and atelectasis may persist.27,28Reopened units recollapse rapidly after discontinuation of PEEP. However, Rothen et al.  29,30demonstrated in volunteers that peak inspiratory pressures of at least 40 cm H²O were needed to fully reverse anesthesia-induced collapse of healthy lungs, and most of the reexpanded atelectatic lung tissue remains inflated for at least 40 min.

Using a unique in vivo  microscopic technique, Halter et al.  31viewed alveoli in the living animal in real time during mechanical ventilation. In a surfactant deactivation model of acute respiratory distress syndrome (ARDS), they measured alveolar number and stability before, during, and after a recruitment maneuver with ventilation with either decreased or increased PEEP. They demonstrated that a recruitment maneuver opens atelectatic alveoli and that without adequate PEEP, alveoli are unstable and susceptible to derecruitment. Although associated with sampling limitations,32this elegant study was the first to directly demonstrate visual evidence of collapse of individual alveoli reflecting atelectasis and stabilization reflecting recruitment.32 

Mechanical ventilation of areas of lung that are atelectatic is associated with repetitive collapse and reexpansion with each breath, often called cyclic recruitment . A conventional view of the effects of atelectasis on gas exchange conveys an image of mixed venous blood perfusing nonventilated or collapsed alveoli, resulting in a consistent and stable proportion of the pulmonary venous return that derives from deoxygenated blood.33This notional contribution (i.e. , deoxygenated blood added to oxygenated blood) is seen as the sum of two constant flow rates, resulting in a shunt fraction (Q^s/Q^t%)7that can be mathematically calculated.

However, we now know that the oxygen content of pulmonary venous blood varies significantly throughout the respiratory cycle as the shunt fraction changes with every breath. Baumgardner et al.  34tested the hypothesis that cyclic collapse of alveoli in dependent lung regions with every breath should lead to large oscillations in Pao²as shunt varies with the respiratory cycle. They placed a partial pressure of oxygen (Po²) probe in the brachiocephalic artery of rabbits after saline lavage. They found a significant effect of respiratory rate on the magnitude of oscillations; as the respiratory rate increased, the amplitude of oscillations became progressively smaller.32The higher respiratory rate avoided cyclic recruitment; i.e. , the shorter expiratory time allowed expiration without collapse. In practice, the beneficial effect of higher respiratory rates leading to improved oxygenation can be seen when ventilating pediatric patients and in particular neonates. This suggests that the dynamics of mechanical events leading to recruitment and derecruitment are important determinants of the amount of lung that is cyclically recruited.34A previous study by Williams et al.  35also showed within-breath arterial Po²oscillations in an animal model of ARDS and confirmed that Pao²oscillations occur in the atelectatic lung. The authors investigated the effect of PEEP on Pao²oscillations. They found that as PEEP was increased, the amplitude of the Pao²oscillation decreased and the mean Pao²increased.35Several studies have evaluated the effects of various inspiratory-to-expiratory ratios on gas exchange, lung mechanics, and FRC; the inspiratory-to-expiratory ratio per se  was not found to be a significant factor.36–38 

More recently, it has been argued that the dependent injured lung is derecruited, not because it is collapsed but because it is filled with fluid.39Work by Gattinoni et al.  40–42had concluded that dependent portions of the injured lung are exposed to a compressive pressure from the increased weight of edematous lung and are collapsed. Such a “weight-of-the-lung” hypothesis has been studied, and it seems that the weight of the lung accounts for only approximately 20% of the vertical gradient in pleural pressure and alveolar volume in normal lungs.43–45Hubmayr39suggests an alternative view of the mechanics of injured lungs based on lung weight, shape matching, interpretation of CT images of the lung, and pressure–volume curves. For example, where edema fluid and foam fill dependent regions, high inflation pressures are required to inflate with air, as opposed to low pressures required for inflation with fluid. The pressure–volume characteristics of the former more closely display the classic findings associated with atelectasis, including the prominent “lower inflection point” reflecting opening pressure.39Evidence in support of edema as the source of regional impedance in ARDS include parenchymal marker studies in oleic acid–injured lungs.46The parenchymal marker technique describes the topographic distribution of regional volume and ventilation in laboratory animals. The authors found that oleic acid injury did not produce collapse of dependent lung units in this model of ARDS.46They proposed an alternative mechanism for the topographic variability in regional impedances and lung expansion after injury, which was liquid or foam in alveoli and conducting airways.46 

It is important for clinicians to understand how atelectasis develops or worsens in the clinical context. Several important clinical events act as modifiers that influence the formation of atelectasis (fig. 2).

Atelectasis develops with both intravenous and inhalational anesthesia, regardless of whether the patient is breathing spontaneously or is paralyzed and mechanically ventilated.47–49Ketamine is the only anesthetic that does not produce atelectasis when used alone, although in conjunction with neuromuscular blockade, it does result in atelectasis.50 

Ventilatory effects of regional anesthesia depend on the type and extension of motor blockade.4Neuroaxial blockade that has significant cephalad extension reduces inspiratory capacity by up to 20%, and expiratory reserve volume approaches zero51; less extensive blockade affects pulmonary gas exchange only minimally, and arterial oxygenation and carbon dioxide elimination are well maintained during most spinal and epidural anesthesia.52,53Closing capacity and FRC remain unchanged.54 

The maximum decrease in FRC seems to occur within the first few minutes of general anesthesia.5,55During anesthesia for surgical operations on the limb, FRC is not influenced further by depth or duration of anesthesia.55,56Don et al.  55studied patients undergoing peripheral limb surgery who were free of cardiac and respiratory disease. Patients were divided into two groups, those breathing halothane in 100% oxygen and those breathing halothane in 30% oxygen in nitrogen. The FRC decreased comparably in both groups after induction of anesthesia, but this did not progress with time. However, other studies have found that during abdominal or thoracic surgery, pulmonary gas exchange deteriorates progressively during the course of the operation57,58; these studies were unable to determine the independent impact of time on atelectasis as opposed to surgical manipulation (e.g. , surgical packing, tissue retraction).

In adults, changing from the upright to the supine position results in a decrease of 0.5 l in FRC to 1.0 l, even in the awake state.7After anesthesia, FRC is reduced by a further 0.5 l to 0.7 l.49The Trendelenburg position allows the abdominal contents to push the diaphragm further cephalad, resulting in a further decrease in FRC.59In the lateral decubitus position, the dependent lung is predisposed to atelectasis, whereas the nondependent lung may have an increased FRC. The overall result is usually a slight increase in total lung FRC, which, despite the differences in lung size, is independent of whether the right or the left lung is in the dependent position.60The prone position may increase FRC slightly,61although this may not decrease atelectasis. In fact, distribution of ventilation is more uniform in anesthetized patients in the prone position, in particular where the abdomen is not supported.62Prone positioning improves oxygenation in patients with ARDS.63Animal models have also demonstrated the benefit of prone positioning after oleic acid–induced lung injury as well as a model of lung injury induced solely by mechanical forces.64,65Prone positioning causes less extensive histologic injury and alters its distribution.64,65 

Atelectasis is more prominent after cardiac surgery with cardiopulmonary bypass (CPB) than after other forms of surgery. Using a porcine model, Magnusson et al.  66found that atelectasis is produced to a much larger extent after CPB than after anesthesia alone or with sternotomy. Furthermore, the CPB-associated atelectasis accounted for most of the marked post-CPB increase in shunt and hypoxemia.66Clinical experience is consistent with laboratory reports, and prominent atelectasis has been noted in the dorsal lung regions on the first postoperative day in cardiac surgery patients.67Other causes of gas exchange impairment after sternotomy and CPB have been investigated, including pulmonary endothelial permeability.68Macnaughton et al.  68did not find any increase in pulmonary endothelial permeability, and they hypothesized that the major component of the deterioration in lung function was probably atelectasis occurring during bypass. Numerous studies have shown that a lung recruitment strategy and PEEP improves oxygenation in patients after CPB.69,70In addition, the application of PEEP of 10 cm H²O during CPB to maintain lung inflation has been shown to be beneficial.71Many anesthetists intentionally inflate the patient’s lungs before coming off CPB and directly visualize equal expansion of the lungs. This simple maneuver may detect significant obstruction from secretions or blood in the endotracheal tube.

High oxygen concentration has been associated with atelectasis formation, and this is important because use of high Fio²(i.e. , approaching 1.0) represents standard practice among many anesthesiologists.72Previous studies suggested that inspired oxygen concentration may not be an important independent predictor of anesthesia-associated atelectasis.5,55However, one of these studies used the helium-dilution technique to measure FRC, which may not be as sensitive as the CT that was used in later studies.55,73In the absence of preoxygenation, atelectasis is not seen directly after induction of anesthesia; however, when Fio²is increased to 1.0 before intubation, development of atelectasis is a consistent finding.74,75Concerns about oxygen-induced atelectasis are not restricted to induction of anesthesia; increasing Fio²at the end of surgery to 1.0 before extubation also causes additional atelectasis, which persists into the postoperative period.76However, the use of lower Fio²may increase the risk of hypoxemia, should airway management subsequently prove difficult and ventilation be threatened.77One study investigated how different oxygen concentrations may affect the formation of atelectasis and the decrease in arterial oxygen saturation during apnea.78The authors found that during routine induction of general anesthesia, 80% oxygen caused minimal atelectasis, but the time margin before desaturation occurred was significantly shortened compared with that of 100% oxygen.78In addition, evidence suggests that administration of supplemental oxygen reduces the incidence of wound infection and could be beneficial during anesthesia and recovery.79,80Another possible benefit of supplemental oxygen during anesthesia includes a reduced incidence of postoperative nausea and vomiting after colorectal surgery,81although this was not found in patients undergoing gynecologic laparoscopy.82 

The effects of nitrous oxide on lung volumes during anesthesia have also been studied.83–85There was no difference in the incidence of postoperative atelectasis if nitrous oxide in oxygen was used or if air in oxygen was used as the inspired gas.

In treating patients for induction of general anesthesia, anesthesiologists must trade off the very rare possibility of acute hypoxemia in the event of difficulty with airway management versus  the common and predictable—but generally mild—impact of hyperoxia-induced atelectasis on later intraoperative gas exchange. Therefore, the use of a lower Fio²to replace preoxygenation has not been recommended as a new standard in clinical practice.78 

In adulthood, progressive age is not associated with increased propensity for development of atelectasis.86However, in young children (aged 1–3 yr) atelectasis seems to develop more readily than in adults,87possibly because of the far greater thoracic wall compliance resulting in less outwardly directed lung distension forces.88In infants, contraction of the diaphragm may cause paradoxical inward movement of the highly deformable chest wall, resulting in loss of lung traction. The resultant atelectasis could reduce ventilatory efficiency, increase diaphragmatic fatigue, and thereby further increase the tendency for atelectasis development. In addition, type I and II muscle fibers are not fully developed in children who are younger than 2 yr. This makes them prone to respiratory failure and fatigue when there is extra stress put on their respiratory system, not only during general anesthesia89but also in the presence of a respiratory tract infection, epiglottitis, or airway obstruction.

Closing volume is also greater in young children in whom the elastic supporting structure of the lung is incompletely developed. This puts the infant at greater risk for atelectasis because airway closure can occur even during tidal breathing.24The closing volume plus the residual volume constitutes the closing capacity. If FRC is decreased relative to closing capacity, this converts normal areas of lung to low V^A/Q areas and compounds the propensity for development of atelectasis.90 

Obesity worsens arterial oxygenation through multiple mechanisms,91of which development of atelectasis is an important contributor. This is because of a markedly reduced FRC promoting airway closure to a greater extent than in healthy-weight subjects.92As the weight of the torso and abdomen make diaphragmatic excursions more difficult—especially when recumbent or supine—the FRC decreases, and this is intensified in the setting of diaphragmatic paralysis associated with neuromuscular blockade. Don et al.  55further clarified the effect of body size on gas exchange and showed that an individual with an increased weight/height ratio (i.e. , obese) would be at a particular disadvantage in the supine position because closing volume is increased in relation to FRC. Pelosi et al.  93also investigated the effects of body mass index on FRC, respiratory mechanics, and gas exchange during general anesthesia. They found that with increasing body mass index (i.e. , ratio of body weight/surface area), FRC and lung compliance decreased, and the oxygenation index (Pao²/Pao²) decreased exponentially.93Prone positioning, although difficult in obese patients, may improve oxygenation.63Consistent with the effects of obesity, the reduced FRC that occurs during pregnancy also potentiates atelectasis.94 

The ARDSnet study recommends the use of low tidal volume in patients with ARDS.95Lower tidal volume reduces stretch-induced lung injury in patients with ARDS,96and this approach is translated into improved patient survival.95,97However, extrapolation of the low-tidal-volume approach to patients without lung injury (or ARDS) requires caution for two principal reasons. First, it has been long recognized that low tidal volume increases the development of atelectasis in the absence of lung injury.2,98–100Therefore, generalized adoption of low tidal volume in patients without preexisting lung injury (e.g. , during general anesthesia) could promote development of atelectasis. Second, in the presence of ARDS, the ARDSnet protocol is specific: The data stipulate 6 ml/kg (vs.  12 ml/kg) in the context of a precise ventilatory management protocol that stipulates many ventilatory parameters, in addition to tidal volume.95A recent study has suggested that the specific “low-tidal-volume” approach used in the ARDSnet study95was actually associated with the development of intrinsic PEEP,97raising the possibility that recruitment—not simply low tidal stretch—played a protective role in these patients. Therefore, if a low tidal volume strategy is used that does not result in the development of intrinsic PEEP, the propensity for development of atelectasis might be significant. The use of low tidal volume in the absence of recruitment or PEEP in anesthetized patients without lung injury may lead to atelectasis and is not recommended.

Two practical points are important regarding mode of ventilation during general anesthesia. In contrast to volume-controlled ventilation, pressure-controlled ventilation results in smaller delivered tidal volumes when respiratory system compliance is decreased (e.g. , during surgical retraction or in the presence of abdominal packs). Smaller tidal volumes may lead to atelectasis and may go undiagnosed because there is no change in the peak airway pressure as pressure-controlled ventilation is being used. In addition, fresh gas flow may have an impact because with older volume-controlled ventilators, increased fresh gas flow results in increased delivered tidal volume.101 

Smokers and patients with lung disease show more pronounced gas exchange impairment in the awake state than healthy subjects do, and this difference also persists during anesthesia.102However, only a small shunt and almost no atelectasis develops in these patients, but they may have severe V^A/Q mismatch. Hyperinflation of the lungs may make them resist collapse.103The chronic hyperinflated state of the lungs in chronic bronchitis changes the mechanical behavior of the lungs and the interaction with the chest wall so that the tendency to collapse is reduced. However, patients with chronic obstructive lung disease may have large regions with low V^A/Q ratios that can result, over time, in resorption atelectasis.4 

Development of atelectasis is associated with the development of several pathophysiologic effects, including decreased compliance, impairment of oxygenation, increased pulmonary vascular resistance, and development of lung injury.

One of the first articles examining the effects of atelectasis was by Mead and Collier99in 1959. They noted that when anesthetized dogs were allowed to breathe spontaneously or were paralyzed and ventilated at tidal volumes of approximately 12.5 ml/kg, pulmonary compliance decreased progressively. They found that these changes were immediately reversed after forced inflations of the lungs, whereas forced deflations caused further compliance reductions. The appearance of the lungs as well as measurement of total and ventilatory lung volumes indicated that the lungs were atelectatic.99 

These laboratory findings were translated into the perioperative context with the article by Bendixen et al. ,2which reported that atelectasis caused a decrease in pulmonary compliance in surgical patients and was associated with a worsening in systemic oxygenation. The decreased compliance is conventionally considered to be due to a reduction in lung volume, such that inspiration–expiration cycles commencing from a lower FRC are completed on a less efficient section of the notional pressure–volume curve.104Therefore, greater energy is consumed because a given change in transpulmonary pressure results in a lesser tidal volume because of the sigmoid shape of the pressure–volume curve. An anatomical basis has been suggested for such atelectasis-associated decreased FRC in the context of general anesthesia (see Compression Atelectasis).6 

The pressure–volume characteristics of the lung determine the work of breathing, and ventilatory work may be analyzed by plotting pressure against volume.105In the presence of increased airway resistance or decreased lung compliance, increased transpulmonary pressure is required to achieve a given tidal volume with consequent increase in the work of breathing.

Although the concept of altered compliance has been well established in the setting of “macroatelectasis” (e.g. , lobar or dependent atelectasis sufficiently large to be radiologically apparent), there are no data regarding whether microatelectasis (e.g. , after brief exposure to increased Fio²during preoxygenation75) has comparable effects on lung mechanics.

In many situations, the most striking effect of atelectasis is impairment of systemic oxygenation. This was first identified in the context of general anesthesia2where the use of passive hyperinflations reversed the hypoxemia.2Others reported impairment of oxygenation during prolonged constant-volume ventilation using small tidal volumes in the absence of intermittent hyperinflations.106–108 

Atelectasis can occur as a result of hyperoxia.75,109If high Fio²is continued, the effects may not be noted by observing pulse oximetry, but only if Pao²is measured. In this situation, the impaired oxygenation can be expressed in terms of Pao². The cause of the impaired oxygenation has been shown, using the multiple inert gas technique, to result from increased mismatch of ventilation with perfusion.75 

Two approaches have been shown to mitigate against the development of hypoxia. First, as demonstrated originally, intermittent hyperinflation maneuvers reverse the effect on gas exchange.2Maintenance of lung volume from the outset can prevent, as opposed to reverse, development of intraoperative atelectasis.28Second, although recognized as an issue for decades, the composition of the inspired gases used during induction of general anesthesia has been shown to directly impact the development of microatelectasis in healthy patients undergoing general anesthesia,75as well as in patients with lung injury.110Rothen et al.  75found that very little atelectasis occurred after the induction of general anesthesia in subjects whose lungs were ventilated with 30% oxygen in nitrogen. Furthermore, the increase in the amount of atelectasis was less rapid in the setting of 30% oxygen (in nitrogen) compared with 100% oxygen. This is consistent with the notion that a poorly absorbed gas such as nitrogen might prevent the early formation of atelectasis and, conversely, that use of a highly absorbed gas (e.g. , oxygen) would enhance the development of atelectasis.

Early studies suggested that the pulmonary vascular resistance was minimal at FRC. Lung volume much above this notional value resulted in alveolar compression due to lung stretch, whereas lung volume falling below the FRC was thought to result in compression of extraalveolar vessels. Therefore, this notion explained the changes in pulmonary vascular resistance on the basis of physical alteration of the pulmonary blood vessels, either stretching or narrowing caused by increased lung volume or compression caused by decreased lung volume.111–113However, regional hypoxia develops in atelectatic lungs, and it has been shown that the mechanism of increased large vessel pulmonary vascular resistance in the lungs is due to hypoxic pulmonary vasoconstriction, due in turn to decreased alveolar and mixed venous oxygen tension.114–116Recent work by our group has demonstrated that this significant increase in vascular resistance can occur in previously normal lungs in an experimental setting, resulting in right ventricular dysfunction and increased microvascular leakage.117 

Extensive evidence has established the importance of maintenance of lung volume in the prevention of lung injury. The pivotal publication by Webb and Tierney118demonstrated that application of PEEP prevented the development of lung injury induced by extremely high tidal volume. The specific degree of atelectasis responsible for attenuation or prevention of high tidal volume–induced lung injury was explored by Sandhar et al.  119and Muscedere et al.  120using different experimental models over 5- and 2-h time periods, respectively. In isolated, nonperfused lungs that are ventilated with no or low PEEP, reductions in compliance and evidence of morphologically apparent lung injury occur. Permitting such atelectasis in the presence of high tidal volumes is associated with hyaline membrane formation, along with regional inhomogeneity of atelectasis and overdistension. Other articles have also demonstrated that repetitive lung collapse or atelectasis leads to increased neutrophil activation in previously injured lungs.121,122In addition to lung injury induced by extremely high tidal volumes, ventilation at low lung volumes worsens lung injury by repeated small airway closure.123,124The concept of permissive hypercapnia developed to avoid lung injury caused by high tidal volumes or high inflation pressures. Higher end-tidal carbon dioxide levels may be accepted to reduce volutrauma or barotrauma, particularly in pediatric patients.

Potentiation of lung injury by atelectasis has additional implications for inflammatory effects in the lung. Tremblay et al.  125examined the effect of ventilation strategy on lung inflammatory mediators in the presence and absence of lung injury and demonstrated that in the absence of PEEP, an impairment in lung compliance was accompanied by increased cytokine concentrations (e.g. , tumor necrosis factor α, interleukin 1β) that were greatest in the groups pretreated with lipopolysaccharide.125These concepts were extended to the in vivo  setting, where it was reported that maximal serum cytokine concentrations induced by high tidal volume occurred in the context of zero PEEP. In addition, atelectasis worsened the impairment of compliance induced by high tidal volume.126Perhaps of greatest concern in that publication was the high mortality (in the absence of cytokine alterations) observed in the in vivo  combination of low-tidal-volume ventilation in the presence of atelectasis. These data suggest that in terms of the most meaningful outcome (i.e. , survival), the combination of low tidal volume and atelectasis may be the most adverse ventilation strategy and that the increased cytokines observed in the high-tidal volume groups may be of less importance.126 

Recent work by our group, wherein rats without lung injury received ventilatory strategies with and without PEEP and recruitment maneuvers, may place these findings into perspective.117We found decreased survival in the zero-PEEP group and also demonstrated ultrastructural evidence of microvascular epithelial disruption in those animals. In addition, we demonstrated that right ventricular dysfunction, possibly secondary to increased pulmonary vascular resistance, occurred in the presence of reduced FRC.117 

Development of atelectasis intraoperatively is associated with decreased lung compliance, impairment of oxygenation, increased pulmonary vascular resistance, and development of lung injury. The adverse effects of atelectasis persist in the postoperative period and can impact patient recovery. Atelectasis can persist for 2 days after major surgery.127Often, the lung dysfunction is transient, and normal lung function resumes soon after anesthesia and surgery. For example, atelectasis resolves within 24 h after laparoscopy in nonobese subjects.128Nevertheless, patients experience perioperative respiratory complications that may be related to reduction in FRC.129,130 

Some pulmonary complications occur during or immediately after anesthesia—mainly hypoxemia. In a large study with more than 24,000 patients, 0.9% had a hypoxemic event in the postanesthesia care unit that necessitated a specific intervention other than only supplemental oxygen.131There is no clear evidence that atelectasis is the cause of all of these postoperative hypoxemic events; respiratory depression from residual anesthetic may be more likely.24However, it seems likely that preventing atelectasis formation during the whole perioperative period would increase the pulmonary oxygen reserve, potentially reducing the likelihood of late postoperative complications.

The characteristic postoperative mechanical respiratory abnormality after abdominal or thoracic surgery is a restrictive pattern with severely reduced inspiratory capacity, vital capacity, and FRC.132–134Patients breathe rapidly with small tidal volumes and are unwilling—or unable—to inspire deeply. Patients in whom postoperative pulmonary complications develop have a relatively greater reduction of FRC, vital capacity, and Pao²than those who do not.132,134The impact of postoperative pain control in preventing postoperative atelectasis has been the focus of much research effort. Although pain and muscle splinting in response to pain are traditionally assumed to be the principal causative factors, total relief of pain after upper abdominal surgery—using epidural analgesia—results in only partial restoration of vital capacity and has minimal effect on FRC.135However, a recent meta-analysis suggested that postoperative epidural opioids significantly decrease the frequency of atelectasis but not other pulmonary complications when compared with systemic opioids.136Other studies have found no difference in the incidence of pulmonary complications. Manikian et al.  137demonstrated that thoracic epidural analgesia in 13 patients undergoing upper abdominal surgery caused a significant increase in forced vital capacity, but the FRC remained unchanged. Spence and Logan138reported no significant impact of postoperative epidural analgesia on oxygenation when compared with patients receiving systemic morphine, and Jayr et al.  139reported that although the epidural provided superior postoperative analgesia, it did not affect the frequency of postoperative pulmonary complications, regardless of preoperative pulmonary status. Finally, a recent multicenter randomized trial of 915 high-risk surgical patients reported a slight reduction in postoperative pulmonary complications—but no impact on mortality or major morbidity—attributable to epidural versus  systemic analgesia, although this study may have been underpowered to examine mortality.140 

Two particular aspects of pulmonary defense mechanisms, coughing and removal of particulate matter, are adversely affected by the changes in lung mechanics and breathing pattern135; this may predispose to pulmonary infection. A number of studies examined the effects of halothane and thiopentone on mucociliary clearance and found that these agents may be responsible for reduced mucous clearance in the postoperative period.141–143A more recent study demonstrated that total intravenous anesthesia with propofol, alfentanil, and vecuronium depressed mucociliary flow in patients with healthy lungs.144Cervin and Lindberg145examined the effects of desflurane on mucociliary activity in the rabbit maxillary sinus in vivo . Desflurane increased mucociliary activity, and the authors concluded that desflurane is an airway irritating compound.145Atelectasis and pneumonia are often considered together because the changes associated with atelectasis may predispose to pneumonia.146Despite a direct lack of evidence of a correlation between atelectasis and pneumonia, reducing or avoiding atelectasis may diminish postoperative pulmonary complications147and thus improve outcome; however, this remains to be proven.

In a series of adults undergoing elective abdominal surgery, postoperative pulmonary complications occurred in 9.6% of cases.148Atelectasis and infectious complications account for the majority of reported pulmonary complications,149and the consequences include a significant burden in terms of morbidity149and additional healthcare costs.150In some series, pulmonary complications account for 24% of deaths within 6 days of surgery,130,149although the precise relation to atelectasis is unclear.

The majority of studies examining the interactions of mechanical ventilation and nonventilatory lung injury have observed the effects of injurious mechanical ventilation strategy on preinjured lungs. However, several experimental studies have examined the effect of preemptive recruitment on the effects of subsequent lung injury.

Multiple studies have convincingly demonstrated that recruitment provides effective protection against lung injury that is either induced118,151,152or aggravated120,125,153,154by mechanical stretch. This may assume increasing importance as the role of tidal volume and plateau pressures in ARDS is debated.95,155,156 

Most patients with severe lung injury require supportive mechanical ventilation. It is clear from multiple laboratory157,158and clinical95,159studies that stretch can cause or worsen lung injury. However, there is a vast spectrum of causes of ALI, and there are striking similarities—and few differences—in the pathogenesis, pathophysiologic dysfunction, and histologic appearance of ALI or ARDS, whether caused by stretch158or other etiologies.160Therefore, because recruitment is effective in reducing stretch-induced injury and because so many pathogenic and morphologic features are shared between stretch-induced versus  other forms of injury, it may be that the protective effects of recruitment could be applicable to lung injury that is not caused by increased tidal stretch.

Additional specific lines of evidence exist. Inflation of a lung graft before reimplantation, as opposed to maintenance of the lung in a deflated state, confers increased viability.161–163The importance of inflation in lung protection has been confirmed in in vivo  162,163and ex vivo  161models and may be mediated either via  reduction in pulmonary vascular resistance161(potentially decreasing endovascular shear injury) or through prevention of surfactant inactivation.163Finally, early application of mechanical ventilation may, through maintenance of lung volume (FRC), reduce mortality in experimental porcine sepsis164and in experimental hemorrhagic shock.165 

Atelectasis is usually suspected when alterations in lung physiology consistent with the development of atelectasis (e.g. , decreased compliance, impaired oxygenation) occur in a setting where atelectasis is likely. However, confirmation of atelectasis is possible through a variety of means.

Lobar or segmental atelectasis is classically represented as opacification of the lobe or lobar segment in question. General signs of atelectasis relate to volume loss. The most direct and reliable sign is the displacement of the interlobar fissure. Other signs of volume loss, such as increase of the hemidiaphragm and mediastinal shift, are maximal nearest the point of volume loss. Compensatory overinflation of the remaining aerated segments in the affected lobe is present, and the collapsed portion of the lung is of increased opacity and often triangular in at least one projection.166For example, when atelectasis results from proximal bronchial occlusion, an obstruction may be identifiable in the proximal bronchial tree, beyond which the tree is not aerated. If atelectasis results from absorption, the features are similar to consolidation, where the atelectatic lung parenchyma is opacified, and contrasts with the patent bronchial airway.

In addition to airway characteristics, other features of atelectasis are important. The “silhouette” sign allows identification of the lobe or segment of the lung that is affected. It is based on the principle that apposition of densely atelectatic lung with an additional contiguous structure, such as the diaphragm or the heart, results in obliteration of the boundary between the lung and the adjacent structure. For example, opacification of part of an atelectatic lung in conjunction with obliteration of the ipsilateral hemidiaphragm suggests lower lobar atelectasis; in contrast, preservation of the hemidiaphragm indicates that the ipsilateral lower lobe is not atelectatic.

A cardinal characteristic of atelectasis is volume loss of the affected lobe. This is associated with secondary changes in adjacent structures, in an attempt to “fill the gap” created by the loss of lung volume, and results in alterations including shift of the mediastinum or the hilum towards the affected area, elevation of the ipsilateral hemidiaphragm, and secondary (or “compensatory”) emphysema in the adjacent nonatelectatic lung.

There is no doubt that conventional chest radiographs can reveal collapse in a segmental or lobar distribution. However, the ability of chest radiographs to detect atelectasis that occurs during general anesthesia or during mechanical ventilation of recumbent critically ill patients is less certain.167 

Computed tomography is emerging as the preferred method for imaging the lung because of the widespread availability, resolution, high signal/noise ratio for lung tissue, and speed. From conventional CT images, it is possible to measure whole and regional lung volumes, distribution of lung aeration, and recruitment behavior under various clinical conditions and interventions.168In the 1980s, atelectasis was shown by CT in anesthetized patients.5,169 

Since then, atelectasis has been studied extensively. Atelectasis on a CT scan has been defined as pixels with attenuation values of −100 to +100 Hounsfield units.3Hounsfield units quantify attenuation or density seen on CT. In this study, Lundquist et al.  3studied patients undergoing elective abdominal surgery with CT of the thorax during anesthesia. Attenuation values in histograms of the lung and atelectasis were studied using two methods of calculating the atelectatic area. On the basis of the CT findings, atelectasis occurs in the most dependent parts of the lungs and occurs in almost 90% of patients who are anesthetized. The extent of the atelectasis in the dependent regions can be reduced by PEEP.5 

The lung in ARDS is characterized by a marked increase in lung tissue and a massive loss of aeration.170Gattinoni et al.  40,41,171,172have extensively investigated the distribution of tidal volume and recruitment in patients with ARDS using CT scans. They concluded that PEEP makes the gas distribution more homogenous in patients with ARDS, stretching the upper levels and recruiting the lower ones, and thereby reduces the atelectatic tissue in dependent regions. Rouby173has assessed PEEP-induced lung overdistension using CT. Rouby determined the threshold to differentiate PEEP-induced alveolar recruitment from PEEP-induced lung overdistension. The threshold of 900 Hounsfield units allows a reliable determination of PEEP-induced alveolar overdistension. A number of human studies have clearly reported the simultaneous onset of alveolar recruitment and lung overinflation in patients receiving PEEP levels of 10 and 20 cm H²O.174–177These recent approaches have important implications in the diagnosis as well as the management of atelectasis because they point clearly to the propensity for injurious regional overinflation occurring as a result of efforts to reverse atelectasis (e.g. , increased PEEP, recruitment maneuvers) in the setting of ARDS.

Magnetic resonance imaging allows three-dimensional imaging without the use of ionizing radiation. Unfortunately, the lung is not well suited to magnetic resonance imaging. On spin-echo images in healthy subjects, little signal is obtained from lung parenchyma. Areas of consolidation and masses can be identified, but the degree to which they can be differentiated has yet to be fully established. This technique has been used in preterm neonates to study pulmonary dysfunction.178The authors compared lung water content and distribution between preterm and term infants and concluded that lung water content is higher in preterm infants, consistent with dependent atelectasis.178Xenon-enhanced CT is a method for noninvasive measurement of regional pulmonary ventilation, determined from the wash-in and wash-out rates of the radiodense, nonradioactive xenon gas, measured by serial CT scans.179This may provide a valuable tool for noninvasive measurement of regional lung function and atelectasis in the future. Although magnetic resonance has several advantages over CT (e.g. , contrast material is not necessary, possibility for multiple planar imaging), it is not apparent that magnetic resonance offers any distinct advantages over CT in the evaluation of atelectasis at the current time.180 

Although more commonly used in assessment of pleural collections or in the context of echocardiography, thoracic ultrasonography has been used in assessment of lung parenchyma.181It has been suggested as a useful aid to the rapid diagnosis of atelectasis182and, in the context of ARDS, seems to be useful for the assessment of regional consolidation.183Recently, use of a specific ultrasonographic detection of the cardiac impulse, termed the lung pulse , has been reported to be a highly sensitive early indicator of the presence of atelectasis.184The use of ultrasonography may offer significant advantages for patients in the operating room or in the intensive care unit in terms of rapidity and ease of examination, without the disadvantages of patient transport and radiation exposure. However, its role in clinical practice remains to be determined.185 

Intravital microscopy applied to the pleural surface enables experimental visualization of atelectasis and examination of its role in the pathogenesis of lung injury (see Direct Visualization of Atelectasis).31In addition to access issues, there are limitations to interpretation,32including the fact that visualization is restricted to lung tissue that is immediately below the pleural surface, with deeper parenchyma beyond the range of the instrument.

There are other methods of detecting atelectasis in the preclinical setting, including lung lavage levels of cytokines.125Tremblay et al.  125examined the effect of ventilation strategy on lung inflammatory mediators in the presence and absence of lung injury using an isolated rat model. Lung lavage cytokine concentrations were greatest in the groups ventilated without PEEP in both the control intravenous lipopolysaccharide–treated groups. Zero PEEP in combination with high volume ventilation had a synergistic effect on cytokine concentrations. This finding that atelectasis worsens stretch-induced lung injury resulting in increased lung and systemic cytokine concentrations was confirmed by Chiumello et al.  126However, it is important to note that experimental atelectasis—even when lethal—is not associated with increased cytokines in the absence of high tidal volumes.126The measurement of lung lavage levels of cytokines in the clinical situation is impractical.

The factors important in the prevention or reversal of atelectasis differ considerably, depending on whether the lungs are injured or uninjured. Because of the adverse pathophysiology associated with the development of atelectasis and the preclinical findings that recruitment of atelectatic lung may reduce the propensity toward subsequent injury,161–163it is important to examine how recruitment may be achieved.

Progressive pulmonary atelectasis (and the associated impairment of oxygenation) may occur during constant ventilation whenever periodic hyperinflation is lacking; it is reversible by passive hyperinflation (i.e. , three successive inflations: first, with a pressure of 20 cm H²O for 10 s; second, with a pressure of 30 cm H²O for 15 s; and third, with a pressure of 40 cm H²O sustained for 15 s).2This was taken as evidence that atelectatic alveoli are reopened by the deep breaths, and that conclusion was supported by the reduction in pulmonary compliance occurring during ventilation and the return of compliance to control values after hyperinflation.

Nunn et al.  107examined arterial oxygenation in patients during routine anesthesia. They found that arterial oxygenation increased when a pressure of 40 cm H²O was maintained for 40 s; lower pressures, even with modest levels of PEEP, were not effective.107A more recent study by Tusman et al.  28examined an “alveolar recruitment” strategy in healthy lungs during general anesthesia. The authors hypothesized that because atelectasis occurs during general anesthesia, an initial increase in pressure would be required to open collapsed alveoli, and if this inspiratory recruitment was combined with sufficient end-expiratory pressure, alveoli would remain open. They allocated patients to one of three groups: no PEEP; an initial control period without PEEP followed by PEEP of 5 cm H²O; and PEEP of 15 cm H²O with high tidal volume (18 ml/kg, or peak inspiratory pressure 40 cm H²O) maintained for 10 breaths, followed by stepwise reduction of PEEP and tidal volume. The third group, receiving the alveolar recruitment strategy, had a significant increase in arterial oxygenation during general anesthesia. Treatment with PEEP of 5 cm H²O alone did not have the same effect on oxygenation. The authors concluded that high initial pressures are needed to overcome the anesthesia-induced collapse and that PEEP of 5 cm or more is required to prevent the newly recruited alveoli from collapsing. In addition, there was no evidence of barotrauma or pulmonary complications as a result of the high initial airway pressures.28 

It has become evident that a number of ventilatory strategies can produce or worsen lung injury.157The use of large tidal volumes,153high peak airway pressures,186and end-expiratory alveolar collapse with cyclic reopening120have all been proposed as deleterious when ventilating injured lungs. Several studies have shown that lung injury secondary to ventilation with large tidal volume is attenuated if end-expiratory volume is maintained by the use of PEEP.118,126,152In addition, ventilation at low FRC worsens lung injury, possibly by repeated small airway closure, and PEEP markedly attenuates this injury.120,126Therefore, recruitment of atelectatic lung reduces the injurious effects of mechanical ventilation with both low and high tidal volumes and protects against development of ventilator-induced lung injury.

The “open-lung” approach to ventilating patients with ARDS consists of a recruitment maneuver with high sustained airway pressures to open atelectatic alveoli followed by application of PEEP to maintain alveolar patency. This approach was among the interventions performed in a study that showed an improvement in oxygenation and reduced mortality in patients with ARDS.159The researchers calculated the inflection point from a pressure–volume curve (corresponding to an upward shift in the slope of the curve), and PEEP was preset at 2 cm H²O above the inflection point in the protective ventilation group. If the inflection point could not be determined on the pressure–volume curve, PEEP of 16 cm H²O was used. Recruiting maneuvers were frequently used and consisted of continuous positive airway pressures of 35–40 cm H²O for 40 s followed by a return to previous PEEP levels.159 

Gattinoni et al.  40also showed that with increasing PEEP from 0 to 20 cm H²O, gas distribution became more homogenous in sedated and paralyzed patients with ARDS and reduced the reopening–collapsing tissue. Suter et al.  187suggested that the optimum end-expiratory pressure in patients with acute respiratory failure was the PEEP level that resulted in the greatest total static compliance. This maximum compliance produced by PEEP resulted in optimum overall cardiopulmonary interaction to ensure maximal global oxygen delivery. In patients with low FRC values at zero end-expiratory pressure, maximum oxygen transport was achieved at higher levels of PEEP than in patients with normal or high FRC values.

The timing of recruitment maneuvers was examined by Grasso et al.  188in patients with ARDS who were ventilated with a lung-protective ventilatory strategy (low tidal volumes of 6 ml/kg). They defined a recruiting maneuver as continuous positive airway pressure of 40 cm H²O for 40 s. Measurements of Pao²/Fio²ratio, volume–pressure curve, and respiratory mechanics were obtained 2 and 20 min after application of a recruiting maneuver. The authors classified the patients as responders and nonresponders. Recruitment maneuvers increased the Pao²/Fio²ratio significantly more in responders—who had been ventilated for a longer period of time—than in nonresponders. The authors concluded that application of recruiting maneuvers improves oxygenation only in patients with early ARDS who do not have impairment of chest wall mechanics and with a large potential for recruitment.188 

Postoperative pulmonary complications are increasingly recognized as an important perianesthetic issue.146Determination of the frequency and clinical impact of postoperative pulmonary complications is hampered by the lack of a uniform definition of a postoperative pulmonary complication among studies. Included in the definition are pneumonia, severe respiratory failure (usually defined as the need for mechanical ventilation), bronchospasm, and atelectasis (often not defined). Most studies referring to postoperative atelectasis do not directly measure FRC.

Reversing or preventing atelectasis is possible in many patients in the perioperative period146,189and is of proven benefit in preventing pulmonary complications.146Techniques or devices that either encourage or force patients to inspire deeply are of most clinical importance.190–192The aim is to produce a large and sustained increase in transpulmonary pressure distending the lung and reexpanding collapsed lung units. Several methods have been studied, including intermittent positive-pressure breathing, deep-breathing exercises, incentive spirometry, and chest physiotherapy. A meta-analysis suggests that all regimens studied are equally efficacious in reducing the frequency of postoperative pulmonary complications after upper abdominal surgery.193 

The beneficial effect on gas exchange of a simple posture change from supine to seated has been demonstrated, both in healthy subjects90and after upper abdominal surgery.91Identification of FRC as the single most important lung volume in postoperative patients provides a specific goal of therapy.189FRC can be measured by a number of techniques. Gas dilution techniques (e.g. , helium equilibration, nitrogen washout) have been used.93,194One disadvantage of these methods is that the tracer gas may not mix with all gas in the lungs because of occluded airways; this source of error can be overcome by using body plethysmography.

Finally, early work reported the use of thoracic or sternal traction195,196as well as the use of intravenous aminophylline197as successful treatments of atelectasis in a number of case reports, but given the impracticalities and adverse effects of these therapies, they have no clinical application today.

Although the pressure–volume relation in ARDS—and the ability to recruit lung—may depend on the etiology,171it is always easier to recruit lung when injury is absent or mild, compared with when significant injury is established.42,188,198This is because the uninjured lung is highly compliant, and its inflection point (opening pressure) on the inflation pressure–volume curve is either low or not detectable. In contrast, the inflection point is significantly greater when injury has been established; in fact, elevation of the inflection point increases in proportion to the extent of injury.194Such impaired compliance (and difficulty with recruitment) has a pathophysiologic basis in established injury. The causes include surfactant dysfunction due to impairment of large aggregate formation,199direct inhibition by plasma proteins that have exudated into the air space,200–202and the progression of cellular infiltration and edema in the pulmonary interstitium and alveoli.203,204Increases in pulmonary vascular resistance, which characteristically parallel the progression of the lung injury,205may decrease lung compliance further.206Finally, when recruitment has been achieved, less airway pressure is required to prevent derecruitment than is required to achieve recruitment.207 

In the future, clinicians may be able to predict the subsequent development of ALI. In an optimal setting, this would be after recognition of the identifiable etiology and before its initiation (e.g. , before major surgery or before reperfusion during transplantation surgery). Alternatively, early institution of prophylactic recruitment after identification of an etiologic event, particularly when associated with a worrisome predictive stratification (e.g. , pancreatitis with elevated cytokine profile), would be ideal. The eventual prospect of predicting (or early detection of) ALI and ARDS makes “prophylactic lung recruitment” (to maintain or recover normal resting lung volume) increasingly tenable, and if validated by clinical studies, this could become integrated into the perioperative care of patients who are at risk of developing critical illness.

Atelectasis occurs in almost all patients undergoing general anesthesia. We have reviewed the causes, effects, nature, identification, and prevention of atelectasis in the perioperative period. Unless restoration of FRC takes place, atelectasis can have detrimental consequences. Current literature indicates that intraoperative and postoperative lung recruitment improves intermediate physiologic outcomes (e.g. , oxygenation, work of breathing); however, the benefits might have more significant implications for lung injury and ARDS. Low stretch ventilation can achieve, at best, attenuation of stretch-induced injury; at worst, it can cause loss of lung volume, potentially contributing to or worsening injury. Prophylactic recruitment however, could reduce or prevent lung injury that results from a variety of primary etiologies, as well as reducing injury resulting from mechanical stretch. If proven in the clinical setting, this approach could reduce the development of critical illness in many general medical and surgical patient populations.

Since acceptance of this manuscript, Squadrone et al . have demonstrated that reversing atelectasis with noninvasive continuous positive airway pressure in high-risk postoperative patients who are hypoxic results in reduced need for reintubation and a lower incidence of pneumonia and sepsis (Squadrone V, Coha M, Cerutti E, Schellino MM, Biolino P, Occella P, Belloni G, Vilianis G, Fiore G, Cavallo F, Ranieri VM; Piedmont Intensive Care Units Network [PICUN]: Continuous positive airway pressure for treatment of postoperative hypoxemia: A randomized controlled trial. JAMA 2005; 293:589–95).

The authors thank Steven Deem, M.D. (Associate Professor, Departments of Anesthesiology and Medicine, University of Washington, Seattle, Washington), Rolf Hubmayr, M.D. (Professor, Department of Medicine, Mayo Clinic, Rochester, Minnesota), and John Laffey, M.D. (Consultant Anesthetist, University Hospital and National University of Ireland, Galway, Ireland), for their expert comments on the manuscript; and Theresa Sakno (Sakno Media, Toronto, Ontario, Canada) for the artwork.