In their classic paper, Bendixen and coworkers proposed “a concept of atelectasis” as a cause of impaired oxygenation during anesthesia [9]. However, atelectasis could not be shown on conventional chest X-ray. With the introduction of radiological computed tomography (CT), densities were seen in dependent lung regions in anesthetized pediatric and adult patients [10, 11]. Morphological studies in various animals showed them to be atelectasis (see an example in Fig. 1.1, left panel).

Atelectasis appears in around 90 % of all patients who are anesthetized during spontaneous as well as mechanical ventilation and whether intravenous or inhalational anesthetics are used [11]. The atelectatic area on a CT near the diaphragm is on average 3–4 % of the total lung area but can easily exceed 15–20 %. The amount of tissue that is collapsed is even larger, the atelectatic area comprising mainly lung tissue, whereas the aerated lung consists of tissue and air. Thus, 10–20 % of the lung is regularly collapsed at the base of the lung during uneventful anesthesia before any surgery has been done. Abdominal surgery does not add much to the atelectasis, but the lung collapse can remain for several days in the postoperative period [12]. After thoracic surgery and cardiopulmonary bypass, more than 50 % of the lung can be collapsed still several hours after surgery [13]. The amount of atelectasis decreases toward the apex that is mostly spared (fully aerated). It is likely that the atelectasis is a locus of infection and that it can contribute to pulmonary complications [14, 15].

The major cause of atelectasis during anesthesia is closure of airways. This is important to remember when considering techniques to prevent atelectasis or reopen collapsed lung tissue. Compression of the lung might be suspected to be a major or additional cause of atelectasis, but this is not likely. Airways will close before alveoli collapse when the lung shrinks. This brings us to the second factor that is needed to cause atelectasis, resorption of the gas that is trapped behind closed airways. The higher the oxygen concentration, the faster is the resorption of gas and atelectasis formation [11] (one may even ask how much of lung collapse in the ARDS patient is caused by compression and how much by gas resorption). Thus, fall in FRC and high oxygen concentration are both needed to produce alveolar collapse, at least when considering the relatively short time of most anesthesias.

Recruitment maneuvers also have been used during cardiac surgery [19] (see also below) and in the intensive care setting [20]. As there is a complex interaction between time and pressure, the time frame possibly differs if other recruitment pressures are used [21]. As an alternative, a stepwise increase in PEEP can be used [22].

Avoiding the preoxygenation procedure and ventilating with 30 % instead of 100 % O₂ prevents formation of atelectasis during the induction and subsequent anesthesia [23]. In studies atelectasis appeared in all patients who were preoxygenated with 100 % O₂, was much smaller with 80 % O₂, and was almost absent with 60 % O₂. However, the smaller amount of atelectasis with lower oxygen concentration during induction remains only for a limited time. The patients receiving 80 % O₂ during induction had as much atelectasis as those on 100 % O₂ 40 min later [24]. This is because the gas trapped behind closed airways consists of 80 % O₂ and will be resorbed during the ensuing period and finally results in airlessness, i.e., atelectasis. Reopening of closed airways by a recruitment maneuver with lower O₂ concentration, e.g., 40 %, even in the absence of atelectasis, will replenish the closed region with lower O₂ gas, and this will slow down resorption atelectasis even more, hopefully for the rest of the anesthesia.

Anesthesia might be induced during ventilation with CPAP that will prevent the fall in FRC and atelectasis formation [25]. Oxygen can be used to full extent, and, moreover, the lung volume is higher compared to no use of CPAP/PEEP, resulting in a larger oxygen reservoir and increased safety time in the event of a complicated intubation of the airway.

The findings of atelectasis during anesthesia and the possibility to recruit lung tissue with an inflation of the lung has prompted studies on the use of recruitment maneuver at the end of the surgery and anesthesia. Again, the influence of inspired oxygen plays an important role. Thus, recruitment at the end of the anesthesia followed by ventilation with 100 % oxygen (the latter again being common in routine anesthesia) caused new atelectasis within the 10 min period before anesthesia was terminated but not if ventilation was with lower FiO₂ [26]. Another approach to prevent atelectasis to persist into the postoperative period is to use PEEP until extubation of the airway and to continue with the CPAP for a limited time, e.g., 15–30 min during which period inspired oxygen concentration is lowered to 30 % in the air. In a small study where this technique was applied, atelectasis was reduced to less than a third compared to control patients with no PEEP/CPAP as assessed by CT one hour after wake up [27].

An individual lung ventilation technique was developed more than 30 years ago in order to optimize ventilation distribution in proportion to individual lung blood flow. It was successful in improving oxygenation but was considered too complicated to be used in intensive care. Thus, it required that:

This made it possible to apply a higher PEEP to the lower lung where most of the atelectasis should be and to ventilate each lung separately so that 50 % of ventilation was given to the upper, nondependent lung and 50 % to the lower, dependent lung. This was assumed to match the distribution of blood flow between the two lungs [28] (Fig. 1.2). Despite its technical complexity, it was also tested in anesthetized patients. Also, during anesthesia, gas exchange could be improved, and CT scanning showed that atelectasis could efficiently be removed from the dependent lung without undue overexpansion of the nondependent lung. The concept has been revived recently, at least experimentally, using better monitoring technique and, more importantly, distributing ventilation in proportion to the lung mechanics of each lung rather than its perfusion. This may not optimize gas exchange but should reduce stress and strain of the lung with possible protective effect on inflammation. Having this as the objective, ventilation will be distributed automatically between the two lungs in proportion to their regional compliances (or, rather, their time constants). Recruitment of collapsed lung tissue with no overexpansion and ventilation of each lung at their optimum PEEP levels can be achieved, as shown in an animal model [29] (see also Fig. 1.3). A simple pneumatic system that allows the use of only one ventilator, still providing different PEEP to the two lungs, does exist [30], and there is today a double lumen tracheal tube that facilitates the insertion and fixation of the tube. The potential value of this in thoracic surgery and in intensive care remains to be studied.

During one-lung ventilation (OLV), one lung is separated from ventilation to enable surgery on that lung, and it does not participate in the pulmonary gas exchange. There is a persisting perfusion that causes shunt and decreased arterial oxygenation. Hypoxic vasoconstriction (HPV) reduces this blood flow, whereas kinking of pulmonary vessels because of compression and distortion of the lung seems to have less effect on blood flow [31].

The other lung is ventilated and perfused. The patient is normally in the lateral position with the non-ventilated lung in the upper position to facilitate surgery and the ventilated lung in the dependent position. Atelectasis is produced in the dependent lung, and pulmonary shunt is regularly larger than 11 % and the PaO₂ reduced by 50 % or more during OLV [32]. A traditional approach to mechanical ventilation during OLV has been high tidal volume (10–12 ml/kg) and zero PEEP to the dependent ventilated lung, high tidal volume to keep the lung open, and no PEEP to preserve the effect of the HPV in the nondependent lung and minimize blood flow to it [33]. However, pulmonary complications are common, both during the anesthesia per se with shunt and hypoxemia (see above) and postoperatively. Pathophysiological disturbances include high airway pressures, ventilation/perfusion mismatch with shunt, increased pulmonary capillary pressure, and cyclic alveolar collapse. These events may result in alveolar damage followed by pulmonary edema with diffuse alveolar injury, leucocyte sequestration, and alveolar cytokine release, a series of events frequently called mechanotransduction (for a review, see [34]). Moreover, on termination of OLV there can be persisting hyperperfusion in the dependent, ventilated lung, associated with an increased diffuse alveolar damage score, as seen in porcine experiments [35].