Cyclical opening and closing of atelectatic alveoli and distal small airways with tidal ventilation is known to be a basic mechanism leading to ventilator-induced lung injury (VILI) 1. To prevent alveolar cycling and derecruitment in acute lung injury (ALI) and acute respiratory distress syndrome (ARDS), high levels of positive end-expiratory pressure (PEEP) have been proposed to counterbalance the increased lung mass resulting from oedema, inflammation and infiltration, and to maintain normal functional residual capacity 2. Although higher levels of PEEP have been shown to prevent VILI in animal studies 13, the random application of either higher or lower levels of PEEP in an unselected population of patients with ALI/ARDS did not significantly improve outcome in three large randomised trials 456. It has been argued that in a partially collapsed lung, high levels of PEEP alone could result in only limited lung protection 4 while exerting its negative effects 78. Therefore, the 'open lung concept' has been proposed 9, aimed at opening up all recruitable alveoli by applying high inflation pressures (lung recruitment manoeuvre (RM) to 'open up the lung'). Once the lung is thought to be recruited, the open lung PEEP (OL-PEEP) is defined as the level of PEEP that prevents end-expiratory collapse ('to keep the lung open'). A decremental PEEP trial after full lung recruitment allows for PEEP titration along the deflation limb of the pressure/volume curve while observing changes in both oxygenation and respiratory mechanics 1011. During a decremental PEEP trial, the point of maximum curvature and maximal tidal respiratory compliance have been shown to correspond to OL-PEEP in theoretical and animal models of ALI/ARDS 101213.

However, high intrathoracic pressures applied during lung recruitment and PEEP titration may cause barotrauma or haemodynamic instability 8141516, representing a potential limitation of the open lung concept. In particular lung recruitment is known to result in significant haemodynamic compromise because of an acute right ventricular pressure overload, with an acute leftward septal shift in transoesophageal echocardiography 141617. On the other hand, re-establishing 'normal' functional residual capacity (FRC) by optimum PEEP should result in unloading of the right ventricle, as pulmonary vascular resistance is related to lung volume in a bimodal fashion, with resistance to flow being minimal near FRC 18. In addition, recruitment of collapsed alveoli, by increasing regional alveolar partial pressure of arterial oxygen (PaO₂), should reduce hypoxic pulmonary vasoconstriction and thus pulmonary vasomotor tone 1920, thereby unloading the right ventricle. Although the potential negative effects of RMs are well defined, it is still unclear whether RMs are beneficial to improve respiratory function when patients with ALI/ARDS are ventilated with high PEEP and low tidal volume, that is using lung protective ventilation.

Therefore, the aims of the present study were to investigate the effects of a standardised, computer-controlled open lung strategy on the respiratory function and haemodynamics in patients with ARDS already being ventilated in a lung protective mode.

This study shows that a standardised open lung strategy consisting of a RM followed by a decremental PEEP trial was effective in improving respiratory system mechanics and oxygenation in patients fulfilling standard ARDS criteria 2127 while already being ventilated with low tidal volume and high PEEP. No clinically significant haemodynamic compromise occurred during the stepwise RM. During the RM, transoesophageal echocardiography revealed increased right ventricular stress and strain, indicated by an increase in right ventricular Tei index, an increase in right ventricular end-diastolic area and a consecutive acute leftward shift of the interventricular septum, resulting in a decreased septal to lateral left ventricular end-diastolic diameter and left ventricular end-diastolic area. During OL-PEEP ventilation, however, right ventricular function assessed by the Tei index was improved compared with baseline conditions with left ventricular function being unchanged.

Two different methods have been proposed as the possible approaches to recruiting the lung: high-level continuous positive airway pressure (CPAP) 2829 and pressure control ventilation with high peak and end-expiratory pressure 30313233. As animal models showed less cardiovascular compromise with the latter approach 34, pressure control ventilation may be considered the optimal approach to lung recruitment 35. Accordingly, in this study we used the pressure control strategy, applying a stepwise increasing peak inspiratory pressure up to 50 cmH₂O at a high level of PEEP, similar to the approach used by Villagra and colleagues 33.

We observed a mean percentage increase in PaO₂/FiO₂ of 22% following the RM and decremental PEEP trial. Furthermore, the improvement in oxygenation was associated with an increase in the dynamic respiratory compliance, suggesting the presence of alveolar recruitment.

The oxygenation response in our study was in line with that reported by Villagra and colleagues 33 but modest compared with the study by Grasso and colleagues 28. This can be explained by different types of patients, the ALI/ARDS onset time and ventilatory setting. In particular, it should be considered that our patients were on a lung protective strategy with low tidal volume and high PEEP (mean PEEP at baseline of 14 cmH₂O), which is likely to result in a lesser improvement in respiratory function after RMs.

The primary complications possibly occurring during RMs are barotrauma and haemodynamic compromise 16173637. RMs may impair haemodynamics, most commonly assessed by MAP or cardiac output, by two main mechanisms 8. First, as the lung is recruited, high airway pressure can more readily be transmitted across the lung parenchyma to the pleural space, impeding venous return and thus decreasing right ventricular preload. Second, high alveolar pressure may increase pulmonary vascular resistance and right ventricular afterload. A recent systematic review 37 revealed hypotension (12%) and desaturation (9%) as the most frequent complications, although serious adverse events such as barotrauma were rare (1%). Given these side effects and the lack of information on the influence on clinical outcome, the authors neither recommend nor discourage RMs at this time. The latter point is especially important, as the effect of RMs is relatively short-lived and RMs must be repeated several times a day in order to maintain open lung ventilation.

The study presented here, albeit small, did not reveal major complications. In particular, we did not observe any significant decrease in MAP, stroke volume or CI during the RMs. Cardiac pumping capability, however, assessed by the cardiac power index, which combines both pressure and flow domains of the cardiovascular system, decreased. These findings of relative haemodynamic stability during the RMs are in line with those reported in the ARDS Network study 438 showing a 10.6 ± 1.2 mmHg decrease in systolic blood pressure during lung recruitment manoeuvre using CPAP over 5 to 10 seconds at 35 to 40 cmH₂O and the study by Borges and colleagues 30 using peak airway pressures up to 60 cmH₂O, where none of the patients investigated experienced haemodynamic compromise during the RMs.

Despite maintained blood pressure and CI, the RMs induced an acute cardiac stress test as evidenced by transoesophageal echocardiography. This implies that monitoring haemodynamics using arterial pressure and cardiac output in clinical practice is likely to miss specific changes in venous return and/or right ventricular loading conditions. Echocardiographic assessment of vena cava diameters, which remained unchanged during the RMs except for maximum IVC diameter, revealed maintained venous return in the present study. The patients in our study were at the lower limits of normovolaemia, as indicated by a mean intrathoracic blood volume index of 883 ml/m² and a stroke volume variation of 14%, suggesting that RMs by pressure control ventilation can safely be performed at low normal volume status without the need to induce potentially detrimental hypervolaemia. The importance of the intravascular volume status during the recruitment manoeuvre has been specifically addressed by Nielsen and colleagues 15 in a porcine lung-lavage model: using transoesophageal echocardiography, they showed left ventricular compromise resulting in a drop in cardiac output during lung recruitment by sustained inflation (40 cmH₂O of CPAP for 30 seconds), which was accentuated by hypovolaemia and attenuated by hypervolaemia. Taken together, these findings underscore the need to ensure an adequate intravascular volume status before attempting RMs.

Although venous return was maintained, the RMs, by inducing lung inflation, most probably increased pulmonary vascular resistance 39, thus increasing right ventricular afterload. This increase in right ventricular afterload could be assessed echocardiographically by the increase in right ventricular Tei index and the increase in right ventricular end-diastolic diameter with a consecutive, acute leftward septal shift, reducing left ventricular size. These findings were not as severe as those seen in the study by Nielsen and colleagues 16, when 40 cmH₂O of CPAP for 10 to 20 seconds was applied to patients following cardiac surgery. Recorded in patients with healthy lungs, these manoeuvres most probably resulted in severe lung overinflation, making the acute right ventricular overload very predictable 1739. The situation may be different in patients with ALI/ARDS, when high airway pressure is less readily transmitted across the lung parenchyma to the pleural space, causing less impairment of venous return and cardiac output 8. This, in addition to the fact that pressure control ventilation instead of sustained inflation was used, may explain the lesser degree of right ventricular dysfunction caused by the RM in the present study.

Although the RM, which is needed as part of the open lung procedure, presents a cardiac stress test mainly due to an acute increase in right ventricular afterload, at OL-PEEP right ventricular function as assessed by the Tei index was even improved compared with baseline settings. Left ventricular function at OL-PEEP was comparable with baseline.

In order to explain these findings, we hypothesise that better oxygenation at lower peak pressure (i.e. better compliance) after a RM and decremental PEEP trial has shifted the ventilation to the deflation limb of the pressure/volume envelope, causing ventilation to take place at higher lung volumes. If this results in higher end-expiratory lung volumes approaching normal FRC, but not causing overdistention, pulmonary vascular resistance will fall due to the U-shaped relation between pulmonary vascular resistance and lung volume. A recent computed tomography study in lung-injured pigs showed that PEEP at which compliance was maximal resulted in the best compromise between recruitment and overinflation 40, which might help to explain the improvement in right ventricular function observed in the present study. These findings are also in keeping with the results from Reis Miranda and colleagues 41, who showed that ventilation according to the open lung concept consisting of high PEEP following a RM did not increase right ventricular outflow impedance compared with conventional ventilation with lower PEEP. The authors propose that resolution of atelectasis due to the RM decreases right ventricular outflow impedance and thus counterbalances the potentially detrimental effects of high PEEP on right ventricular function 8. In fact, Duggan and colleagues showed that atelectasis causes significant increases in right ventricular afterload and that this may even lead to right ventricular failure in healthy rats 42.

To better interpret our results, some limitations need to be addressed. A relatively small number of patients were included in the study due to a selection of more severe patients with early ARDS and absence of haemodynamic instability and without significant arrythmias. As we investigated a specific RM, it is possible that different results could be obtained by using other manoeuvres. Finally, the measurements were made only at the end of the recruitment procedure, which overall lasts for six minutes. The clinical consequence of the RM may not be trivial and in order to keep the lung open the RM must be repeated several times a day in clinical practice.

In conclusion our study demonstrates that standard recruitment manoeuvres during protective ventilation can be associated with haemodynamic changes not revealed by conventional haemodynamic monitoring. A decremental titration of PEEP aimed to yield maximum dynamic compliance was associated with an improvement in oxygenation, dynamic respiratory system compliance and unloading the right ventricle while not affecting the left ventricle.

• A standardised open lung strategy consisting of a recruitment manoeuvre followed by a decremental OL-PEEP trial inproves oxygenation and respiratory system compliance in patients with ARDS already ventilated in a lung protective mode.

• Although major haemodynamic indices remain unchanged, transoesophageal echocardiography reveals increased right ventricular stress and strain during the recruitment phase.

• Compared with baseline values, right ventricular function is improved at OL-PEEP.

ALI: acute lung injury; ARDS: adult respiratory distress syndrome; Cdyn: dynamic compliance of the respiratory system; CI: cardiac index; CPAP: continuous positive airway pressure; EIP: end-inspiratory pressure; FiO₂: fraction of inspired oxygen; FRC: functional residual capacity; IBW: ideal body weight; IVC: inferior vena cava; MAP: mean arterial pressure; OL-PEEP: open lung positive end-expiratory pressure; PaCO₂: partial pressure of arterial carbon dioxide; PaO₂: partial pressure of arterial oxygen; PEEP: positive end-expiratory pressure; PiCCO: Pulse Contour Cardiac Output Monitor; RM: recruitment manoeuvre; RR: respiratory rate; T₀: time at baseline; T_20/30: time when positive end-expiratory pressure at 20 cmH₂O and driving pressure at 30 cmH₂O; T_OLP: time at open lung positive end-expiratory pressure; VILI: ventilator-induced lung injury; Vtinsp: inspiratory tidal volume.

The authors declare that they have no competing interests.

CG, GW, PP and TL participated in the study design. CG, GW and TL performed the study. CG and TL processed the data and performed the statistical analysis. TL and PP wrote the manuscript. All authors read and approved the final manuscript.