In patients with ARDS, increasing the ventilator respiratory rate is the simplest way to enhance CO₂ elimination when tidal volume is reduced 51617. However, an uncontrolled increase in respiratory rate may generate intrinsic PEEP 1819, which, in turn, may promote excessive intrathoracic pressure and lung overinflation 20. If the inspiratory time is not decreased in proportion to the increase in respiratory rate, the resulting intrinsic PEEP may even cause right ventricular function to deteriorate 21. In addition to inappropriate ventilator settings – high respiratory rate together with high inspiratory to expiratory ratio – airflow limitation caused by bronchial injury promotes air trapping 2223. Acting in the opposite direction, external PEEP reduces intrinsic PEEP and provides a more homogeneous alveolar recruitment 2425, whereas lung stiffness tends to accelerate lung emptying 1626. As a consequence, in a given patient, it is impossible to predict intrinsic PEEP induced by a high respiratory rate and no 'magic number' can be recommended. At the bedside, the clinician should increase the ventilator respiratory rate while looking at the expiratory flow displayed on the screen of the ventilator: the highest 'safe respiratory rate' is the rate at which the end of the expiratory flow coincides with the beginning of the inspiratory phase (Fig. 1).

When CO₂ elimination is impaired by tidal volume reduction, the CO₂-laden gas present at end expiration in the physiological dead space is readministered to the patient at the beginning of the following inspiration. The physiological dead space consists of three parts: first, the instrumental dead space, defined as the volume of the ventilator tubing between the Y piece and the distal tip of the endotracheal tube; second, the anatomical dead space, defined as the volume of the patient's tracheobronchial tree from the distal tip of the endotracheal tube; and third, the alveolar dead space, defined as the volume of ventilated and nonperfused lung units. Only the former can be substantially reduced by medical intervention. Prin and colleagues have reported that replacing the heat and moisture exchanger by a conventional heated humidifier positioned on the initial part of the inspiratory limb induces a 15% decrease PaCO₂ in by reducing CO₂ rebreathing 27 (Fig. 2). With a conventional humidifier, the temperature of the inspired gas should be increased at 40°C at the Y piece so as to reach 37°C at the distal tip of the endotracheal tube 27. In sedated patients, the tubing connecting the Y piece to the proximal tip of the endotracheal tube can also be removed to decrease instrumental dead space 16. For the same reason, if a capnograph is to be used, it should be positioned on the expiratory limb, before the Y piece. Richecoeur and colleagues have shown that optimizing mechanical ventilation by selecting the appropriate respiratory rate and minimizing instrumental dead space allows a 28% decrease in PaCO₂ 16 (Fig. 2).

The basic principle of expiratory washout is to replace, with afresh gas, the CO₂-laden gas present at end expiration in theinstrumental dead space 28. It is aimed at further reducing CO₂ rebreathing and PaCO₂ without increasing tidal volume29. In contrast to tracheal gas insufflation, in which the administration of a constant gas flow is continuous over the entire respiratory cycle, gas flow is limited to the expiratory phase during expiratory washout. Fresh gas is insufflated by a gas flow generator synchronized with the expiratory phase of the ventilator at flow rates of 8 to 15 L/min through an intratracheal catheter or, more conveniently, an endotracheal tube positioned 2 cm above the carina and incorporating an internal side port opening in the internal lumen 1 cm above the distal tip 1629. A flow sensor connected to the inspiratory limb of the ventilator gives the signal to interrupt the expiratory washout flow when inspiration starts. At catheter flow rates of more than 10 L/min, turbulence generated at the tip of the catheter enhances distal gas mixing, and a greater portion of the proximal anatomical dead space is flushed clear of CO₂, permitting CO₂ elimination to be optimized 3031. Expiratory washout can be applied either to decrease PaCO₂ while maintaining tidal volume constant or to decrease tidal volume while keeping PaCO₂ constant. In the former strategy, expiratory washout is used to protect pH, whereas in the latter it is used to minimize the stretch forces acting on the lung parenchyma, to minimize ventilator-associated lung injury.

Two potential side effects should be taken into considerationif expiratory washout is used for optimizing CO₂ elimination. Intrinsic PEEP is generated if the expiratory washout flow is not interrupted a few milliseconds before the beginning of the inspiratory phase 1629. As a consequence, inspiratory plateau airway pressure may increase inadvertently, exposing the patient to ventilator-induced lung injury. If expiratory washout is to be used clinically in the future, the software synchronizing the expiratory washout flow should give the possibility of starting and interrupting the flow at different points of the expiratory phase. A second critical issue conditioning the clinical use of expiratory washout is the adequate heating and humidification of the delivered washout gas.

Currently, expiratory washout is still limited to experimental use. It is entering a phase in which overcoming obstacles to clinical implementation may lead to the development of commercial systems included in intensive-care-unit ventilators that may contribute to optimizing CO₂ elimination 30, in particular in patients with severe acute respiratory syndrome associated with head trauma 32.

Periodic increases in inspiratory airway pressure may contribute to the optimization of alveolar recruitment in patients with ALI and ARDS. Sighs are characterized by intermittent increases in peak airway pressure, whereas recruitment manoeuvres are characterized by sustained increases in plateau airway pressures. The beneficial impact of sighs and recruitment manoeuvres on lung recruitment is based on the well-established principle that inspiratory pressures allowing reaeration of the injured lung are higher than the expiratory pressures at which lung aeration vanishes. At a given PEEP, the higher the pressure that is applied to the respiratory system during the preceding inspiration, the greater the lung aeration. In patients with ALI, the different pressure thresholds for lung aeration at inflation and deflation depend on the complex mechanisms regulating the removal of oedema fluid from alveoli and alveolar ducts 3334, the reopening of bronchioles externally compressed by cardiac weight and abdominal pressure 35, and the preservation of surfactant properties.

Reaeration of the injured lung basically occurs during inspiration. The increase in airway pressure displaces the gas – liquid interface from alveolar ducts to alveolar spaces and increases the hydrostatic pressure gradient between the alveolar space and the pulmonary interstitium 36. Under these conditions, liquid is rapidly removed from the alveolar space, thereby increasing alveolar compliance 37 and decreasing the threshold aeration pressure. Surfactant alteration, a hallmark of ALI, results from two different mechanisms: direct destruction resulting from alveolar injury, and indirect inactivation in the distal airways caused by a loss of aeration resulting from external lung compression 38. By preventing expiratory bronchiole collapse, PEEP has been shown to prevent surfactant loss in the airways and avoid collapse of the surface film 38. As a consequence, alveolar compliance increases and the pressure required for alveolar expansion decreases. The time scale for alveolar recruitment and derecruitment is within a few seconds 3940, whereas the time required for fluid transfer from the alveolar space to the pulmonary interstitium is of the order of a few minutes 36. It has been demonstrated that the beneficial effect of recruitment manoeuvres on lung recruitment can be obtained only when the high airway pressure (inspiratory or incremental PEEP) is applied over a sufficient period 4142, probably preserving surfactant properties and increasing alveolar clearance 14.

In surfactant-depleted collapse-prone lungs, recruitment manoeuvres increase arterial oxygenation by boosting the ventilatory cycle onto the deflation limb of the pressure – volume curve 42. However, in different experimental models of lung injury, recruitment manoeuvres do not provide similar beneficial effects 43. In patients with ARDS, recruitment manoeuvres and sighs are effective in improving arterial oxygenation only at low PEEP and small tidal volumes 4445. When PEEP is optimized, recruitment manoeuvres are either poorly effective 46 or deleterious, inducing overinflation of the most compliant lung regions 47 and haemodynamic instability and worsening pulmonary shunt by redistributing pulmonary blood flow towards non-aerated lung regions 48. However, after a recruitment manoeuvre, a sufficient PEEP level is required for preventing end-expiratory alveolar derecruitment 49. Furthermore, recruitment manoeuvres are less effective when ALI/ARDS is due to pneumonia or haemorrhagic oedema 43.

Different types of recruitment manoeuvre have been proposed for enhancing alveolar recruitment and improving arterial oxygenation in the presence of ALI 50. A plateau inspiratory pressure can be maintained at 40 cmH₂O for 40 s. Stepwise increases and decreases in PEEP can be performed while maintaining a constant plateau inspiratory pressure of 40 cmH₂O 42. Pressure-controlled ventilationusing high PEEP and a peak airway pressure of 45 cmH₂O can be applied for 2 min 51. The efficacy and haemodynamic side effects have been compared between three different recruitment manoeuvres in patients and animals with ARDS 4951. Pressure-controlled ventilation with high PEEP seems more effective in terms of oxygenation improvement, whereas a sustained inflation lasting 40 seconds seems more deleterious to cardiac output 4951.

Studies reporting the potential deleterious effects of recruitment manoeuvres on lung injury of regions remaining fully aerated are still lacking. As a consequence, the administration of recruitment manoeuvres should be restricted to individualized clinical decisions aimed at improving arterial oxygenation in patients remaining severely hypoxaemic. As an example, recruitment manoeuvres are quite efficient for rapidly reversing aeration loss resulting from endotracheal suctioning 52 or accidental disconnection from the ventilator. In patients with severe head injury, recruitment manoeuvres may cause cerebral haemodynamics to deteriorate 53. As a consequence, careful monitoring of intracranial pressure should be provided in case of severe hypoxaemia requiring recruitment manoeuvres.

Turning the patient into the prone position restricts the expansion of the cephalic and parasternal lung regions and relieves the cardiac and abdominal compression exerted on the lower lobes. Prone positioning induces a more uniform distribution of gas and tissue along the sternovertebral and cephalocaudal axis by reducing the gas/tissue ratio of the parasternal and cephalic lung regions 5455. It reduces regional ventilation-to-perfusion mismatch, prevents the free expansion of anterior parts of the chest wall, promotes PEEP-induced alveolar recruitment 56, facilitates the drainage of bronchial secretions and potentiates the beneficial effect of recruitment manoeuvres 57, all factors that contribute to improving arterial oxygenation in most patients with early acute respiratory failure 55 and may reduce ventilator-induced lung overinflation.

It is recommended that the ventilatory settings be optimized before the patient is turned into the prone position 35. If arterial saturation remains below 90% at an inspiratory fraction of oxygen of at least 60% and after absolute contraindications such as burns, open wounds of the face or ventral body surface, recent thoracoabdominal surgical incisions, spinal instability, pelvic fractures, life-threatening circulatory shock and increased intracranial pressure have been ruled out 56, the patient should be turned to prone in accordance with a predefined written turning procedure 56. The optimum duration of prone positioning remains uncertain. In clinical practice, the duration of pronation can be maintained for 6 to 12 hours daily and may be safely increased to 24 hours 58. The number of pronations can be adapted to the observed changes in arterial oxygenation after supine repositioning 55. Whether the abdomen should be suspended during the period of prone position is still debated 56. Complications are facial oedema, pressure sores and accidental loss of the endotracheal tube, drains and central venous catheters. Despite its beneficial effects on arterial oxygenation, clinical trials have failed to show an increase in survival rate by prone positioning in patients with acute respiratory failure 5960. Whether it might reduce mortality and limit ventilator-associated pneumonia in the most severely hypoxaemic patients 5960 requires additional study.

Endotracheal suctioning is routinely performed in patients with ALI/ARDS. A negative pressure is generated into the tracheobronchial tree for the removal of bronchial secretions from the distal airways. Two factors contribute to lung derecruitment during endotracheal suctioning: the disconnection of the endotracheal tube from the ventilator and the suctioning procedure itself. Many studies have shown that the sudden discontinuation of PEEP is the predominant factor causing lung derecruitment in patients with ALI 5261. During a suctioning procedure lasting 10 to 30 seconds, the high negative pressure generated into the airways further decreases lung volume 15. A rapid and long-lasting decrease in arterial oxygenation invariably results from open endotracheal suctioning 62. It is caused by a lung derecruitment-induced increase in pulmonary shunt and a reflex bronchoconstriction-induced increase in venous admixture; both factors increase the ventilation/perfusion ratio mismatch 52. The decrease in arterial oxygenation is immediate and continues for more than 15 min despite the reestablishment of the initial positive end-expiratory level. A recruitment manoeuvre performed immediately after the reconnection of the patient to the ventilator allows a rapid recovery of end-expiratory lung volume and arterial oxygenation 62. However, in the most severely hypoxaemic patients the open suctioning procedure itself may be associated with dangerous hypoxaemia 62.

Closed-circuit endotracheal suctioning is generally advocated for preventing arterial oxygenation impairment caused by ventilator disconnection 6364. However, a loss of lung volume may still be observed, resulting from the suctioning procedure itself and appearing dependent on the applied negative pressure 1563. Both experimental studies and clinical experience suggest that closed-circuit endotracheal suctioning is less efficient than open endotracheal suctioning for removing tracheobronchial secretions 6465. As a consequence, the clinician is faced with two opposite goals: preventing lung derecruitment and ensuring the efficient removal of secretions 66. Further clinical studies are needed to evaluate an optimum method that takes both goals into account.

In patients with ALI/ARDS, closed-circuit endotracheal suctioning should be considered the clinical standard. In severe ARDS, endotracheal suctioning should be optimized by pre-suction hyperoxygenation and followed by post-suction recruitment manoeuvres. In addition to the methods described above, two other types of recruitment manoeuvre have been proposed to prevent a loss of lung volume and reverse atelectasis resulting from endotracheal suctioning: the administration of triggered pressure-supported breaths at a peak inspiratory pressure of 40 cmH₂O during suctioning 15 and the administration of 20 consecutive hyperinflations set at twice the baseline tidal volume immediately after suctioning 52.

There is as yet no guideline for endotracheal suctioning in patients with severe ARDS. An algorithm is proposed in Fig. 3 aimed at preventing lung derecruitment and deterioration of gas exchange during endotracheal suctioning in hypox-aemic patients receiving mechanical ventilation with PEEP.