“Open lung” mechanical ventilation is a protective ventilation modality that aims to minimize lung collapse, prevent cyclic alveolar aperture and closure, and reduce pulmonary overdistension.1–3 Its implementation is mainly based on three interventions: (1) the early application of lung recruitment maneuvering (LRM); (2) the use of a positive end-expiratory pressure (PEEP) level sufficient to stabilize and keep open the maximum possible number of alveolar units (the so-called “open-lung PEEP”); and (3) ventilation with as low a pulmonary distension pressure as possible.4,5 However, in order to reach these objectives it is necessary to at least temporarily use very high intrathoracic pressures, and for this reason application of the above ventilatory strategy remains the subject of debate.6,7

The advocators of open lung ventilation justify its use in terms of the harmful effects of lung collapse upon the evolution of the lesions associated to mechanical ventilation, favored by the use of small tidal volumes.8 In contrast, the critics of the technique point out that in addition to the absence of solid confirmation of its benefits in terms of lessened patient mortality, the application of high intrathoracic pressures is not risk-free, with non-negligible hemodynamic alterations and the appearance of barotrauma phenomena.6,9

In September 2003 we introduced an open lung mechanical ventilation protocol based on the use of LRM followed by the application of open-lung PEEP. The aim of this ventilatory strategy was to recruit the maximum possible number of alveoli, ventilating with maximum respiratory system compliance (Crs), and thus with the lowest lung ventilation pressure.2,10 The recent debate published in this journal on the use of this ventilation mode6,7 has prompted us to present some of the most relevant data which we have collected during these years of experience. Specifically, the present retrospective analysis has two aims: (1) to investigate the respiratory and hemodynamic effects of LRM in a group of patients subjected to careful monitoring; and (2) to establish the incidence and form of presentation of barotrauma, and describe the main characteristics and evolution of the patients who develop this complication. While the first of these two points constitutes the main objective of the present study, the second point, related to lung barotrauma, is dealt with in another part of this issue.11 Some of the results of the present study have been previously reported in abstract form.12

The principal patients characteristics are summarized in Table 1. In most cases, the cause of respiratory failure was lung infection (67%). At the time of LRM, a total of 12 patients were receiving vasoactive or inotropic medication (noradrenalin in 9 cases, noradrenalin plus dobutamine in 2, and dobutamine alone in one patient). The mean noradrenalin and dobutamine dose was 0.24±0.09 and 6.91±0.43μgkg−1min−1, respectively. The respiratory frequency and tidal volume at the start of maneuvering was 16.6±2.2rpm and 6.9±1.1ml/kg body ideal weight, respectively.

Effects of recruitment upon respiratory function

The evolution of the main respiratory parameters during LRM is reported in Table 2. Fig. 2A in turn shows the evolution of the respiratory parameters in one of the patients analyzed.

Table 2.

Respiratory changes during pulmonary recruitment maneuvering.

	preRM	maxPEEP	bestPEEP
PEEP, cmH2O	14.1±2.1	35.1±2.1††	16.8±2.7†
Tidal volume, ml	437.6±87.1	231.3±59.9††	488.5±93.2**
Compliance, ml/cmH2O	26.5±7.8	12.4±3.9††	32.4±10.6††
Peak pressure, cmH2O	31.3±3.9	54.2±3.7††	32.8±4.2†
Mean pressure, cmH2O	21.9±2.8	43.9±2.7††	24.3±3.3†
Ventilation pressure, cmH2O	17.2±2.9	19.2±2.9††	15.9±3.4*
SpO2/FiO2, %	101.2±17.7		171.2±57††

Data expressed as the mean±standard deviation.

*P<0.05 vs preMR, **P<0.01 vs preMR, †P<0.001 vs preMR, ††P≤0.0001 vs preMR. bestPEEP: PEEP at maximum respiratory compliance; FiO2: inspiratory oxygen fraction; maxPEEP: maximum PEEP reached during recruitment maneuvering; PEEP: positive end-expiration pressure; PreRM: basal values, before start of recruitment maneuvering; SpO2: oxygen saturation measured by pulsioxymetry.

Figure 2.

(A) Respiratory registry, from bottom to top: peak pressure, PEEP, tidal volume and respiratory compliance. (B) Hemodynamic registry: Accel: acceleration of aortic flow; CO: cardiac output; Diameter: aortic diameter; HR: heart rate; SAF: systolic aortic flow; LVETc: heart rate corrected left ventricle ejection time; MBP: mean blood pressure; SV: systolic volume; SVa: systolic volume, descending aorta; TSVR: total systemic vascular resistance; TSVRa: total vascular resistance of the descending aorta; PV: peak velocity.

(0.54MB).

The maximum PEEP level reached during maneuvering was 35.1±2.1cmH2O, with a maximum peak pressure of 54.2±3.7. The open-lung PEEP value was 16.8±2.7cmH2O, and was higher in obese patients (body mass index>30kg/m2): 19.6cmH2O (17.8–20.4) versus 15.7cmH2O (14.8–17.5; p<0.05). At maximum PEEP level, Crs decreased 52% with respect to the initial value (95%CI: 48.4–56.2%), while at the end of LRM it increased 23±15.2% (95%CI: 15.91–30.14%). In 16 patients (76%) this increase was ≥15% with respect to the basal value (15.3–56.3%), and in only 6 patients was it ≥30% (33.8–56.3%).

In the 15 patients subjected to central venous blood gas analysis, LRM induced or increased hypercapnia with respiratory acidosis during the maximum PEEP phase. After establishing the open-lung PEEP value, these alterations were maintained with respect to the initial values (Table 3), while the SpO2/FiO2 ratio increased significantly (Table 2).

Table 3.

Venous blood gas measurements during pulmonary recruitment maneuvering.

	PreRM	maxPEEP	bestPEEP
pH	7.24±0.09	7.08±0.11†	7.12±0.13*
PvCO2, mmHg	61.6±13.7	95.8±23†	85.8±22.4**
HCO3, mEq/l	25.2±5.6	26.8±5.9*	26.8±5.7*

n=15. Data expressed as the mean±standard deviation.

*P<0.01 vs preRM; **P<0.001 vs preRM; †P≤0.0001 vs preRM. bestPEEP: PEEP at maximum respiratory compliance; maxPEEP: maximum PEEP reached during recruitment maneuvering; PreRM: basal values, before start of recruitment maneuvering; PvCO2: venous CO2 pressure.

Comparison of the ascending and descending branches of recruitment maneuvering

On contrasting comparable PEEP levels in both LRM branches, with similar mean airway pressures, the Crs value was always greater over the descending branch (Fig. 3 and Table 5). From the hemodynamic perspective, the modification of LVETc, SV and CO was more pronounced over the descending branch, while heart rate was greater during this LRM stage (Figs. 4 and 5, and Table 6).

Figure 3.

Evolution of respiratory compliance during lung recruitment maneuvering at comparable levels of PEEP. Each level is represented by its mean value and standard deviation.

(0.18MB).

Table 5.

Respiratory values compared during the ascending and descending branches of recruitment maneuvering at comparable levels of PEEP.

	Ascending branch					Descending branch
PEEP, cmH2O	18.9±2.3	22.9±2.3	27±2.3	30.9±2.5	PEEP maximum	31.2±2.1	27.4±2.1	22.8±2.8	18.9±2.6
Tidal volume, ml	397±79	349±77	296±69	261±62		274±72*	331±85†	409±87†	466±102†
Compliance, ml/cmH2O	23.2±7.3	19.9±6.7	16.4±5.4	14.2±4.7		14.8±5.1	18.4±6.5*	24.5±8.7†	30.1±10.4†
Peak pressure, cmH2O	36.6±4.4	41.1±4.6	45.5±4.4	49.8±4.4		50.2±3.9	45.9±4.1	40.2±4.8	35.1±4.2*
Mean pressure, cmH2O	26.8±3.2	30.9±3.3	35.3±3.1	39.6±3.3		40.1±2.8	35.9±2.9	30.6±3.6	26.3±3.2

Data expressed as the mean±standard deviation.

*P<0.01 decremental vs incremental; **P<0.001 decremental vs incremental; †P≤0.0001 decremental vs incremental.

bestPEEP: PEEP at maximum respiratory compliance; maxPEEP: maximum PEEP reached during recruitment maneuvering; PEEP: positive end-expiratory pressure. Ventilation pressure=peak pressure−PEEP; PreRM: basal values, before start of recruitment maneuvering.

Figure 4.

Evolution of cardiac output during lung recruitment maneuvering at comparable levels of PEEP. Each level is represented by its mean value and standard deviation.

(0.11MB).

Figure 5.

Evolution of heart rate corrected left ventricle ejection time (LVETc) at comparable levels of PEEP. Each level is represented by its mean value and standard deviation.

(0.12MB).

Table 6.

Hemodynamic values compared during the ascending and descending branches of recruitment maneuvering at comparable levels of PEEP.

	Ascending branch					Descending branch
PEEP, cmH2O	18.9±2.3	22.9±2.3	27±2.3	30.9±2.5	PEEP maximum	31.2±2.1	27.4±2.1	22.8±2.8	18.9±2.6
CO, l/min	6.46±2.2	6.29±2.2	5.97±2.1	5.47±1.9		5.06±1.7*	5.17±1.9*	5.71±2.11*	5.91±2.08
SV, ml	62.2±19.9	60.4±18.9	56.8±18.4	51±16.8		44.4±15.5†	45.7±16.7†	49.9±17.9†	51.8±17.8**
HR, bpm	105.3±20.2	105.4±19.9	106.2±19.5	108.5±19.1		112.6±19.4**	114.4±19.9†	115.5±18.3††	115.5±18.8†
MBP, mmHg	81.5±12.5	82.4±12.5	84.7±13.2	87.7±13.3		90.8±11.3	91.5±11.1*	85.9±13.7	80.1±12.2
CW, W	1.14±0.3	1.12±0.3	1.09±0.3	1.05±0.3		1.01±0.3	1.04±0.4	1.07±0.4	1.04±0.4
Accel, ms2	25.1±13.2	26.3±13.9	27.4±15.2	28.5±17.6		30.6±21.4	30.9±21.3	29.5±17.5	28.2±16
LVETc, ms	288.9±45.2	282.1±47.2	272.9±52.4	258.4±53.8		238.8±53.3†	236.9±53.1†	256.1±51.5††	265.2±52.7†
TSVR, dynscm−5	1165±540	1197±509	1310±564	1473±626		1696±856	1717±928*	1477±846*	1277±639

Data expressed as the mean±standard deviation.

*P<0.05 decremental vs incremental, **P<0.01 decremental vs incremental, †P≤0.001 decremental vs incremental, ††P≤0.0001 decremental vs incremental.

Accel: maximum acceleration of aortic flow; bestPEEP: PEEP at maximum respiratory compliance; CW: cardiac power; HR: heart rate; CO: cardiac output; LVETc: corrected left ventricle ejection time; maxPEEP: maximum PEEP reached during recruitment maneuvering; MBP: mean blood pressure; PEEP: positive end-expiratory pressure; PreRM: basal values, before start of recruitment maneuvering; TSVR: total systemic vascular resistance; SV: systolic volume.

In this study, lung recruitment maneuvering (LRM) comprising an ascending branch with progressive (stepwise) PEEP increments followed by a descending branch until reaching maximum respiratory system compliance (Crs) conditioned respiratory changes characterized by an important decrease in Crs at maximum PEEP level and an increase at the end of LRM with values higher than at the start, together with improvement of the SpO2/FiO2 ratio and a decrease in ventilation pressure. The hemodynamic changes fundamentally consisted of a decrease in CO, SV and LVETc, together with an increase in heart rate and contractility. At comparable PEEP levels and mean airway pressure, significant differences were observed between the two LRM branches, with more marked modification of left ventricle preload and CO over the descending branch.

The increase in Crs following LRM and its variations in the form of stepwise PEEP increments and decrements have been demonstrated in different clinical and experimental studies,21–24 and have been carefully detailed by Hickling in a theoretical mathematical model.25 According to the observations of this author, the changes in Crs during this kind of maneuver depend on: (a) the interaction between the number of alveoli that remain stable and inflated over the entire respiratory cycle; (b) the transalveolar pressure and degree of overdistension reached; and (c) the number of initially collapsed alveoli but which become inflated in the course of the inspiratory period (so-called “tidal recruitment”26). In our patients, the changes in Crs during LRM were similar to those predicted by the mathematical model of Hickling, with the exception that in our study the starting PEEP had been previously established according to the maximum Crs, though without prior opening maneuver. For this reason it is probable that in addition to the number of aerated stable alveoli and the transalveolar pressure reached, a non-negligible component of tidal recruitment could influence the Crs value before performing LRM. Likewise, the stepwise decrease in Crs with each PEEP increment after starting the maneuver could have been related to the gradual increase in transalveolar pressure and the successive decrease in tidal recruitment; accordingly, on reaching the upper transalveolar pressure limit, the lung overdistension value was maximum and tidal recruitment minimum or inexistent—this resulting in Crs at its lowest level. Lastly, during the descending phase, the stepwise decrease in transalveolar pressure gave rise to a gradual increase in Crs until reaching its maximum value, which in accordance to previous observations should coincide with the reinitiation of lung collapse.21,27 It is very likely that the explanation for the observed differences in Crs between both branches of the maneuver, at comparable PEEP levels, is the existence of a larger number of permanently aerated alveoli over the descending branch of the maneuver, which would support the efficacy of LRM.21,25

Together with an improvement in the SpO2/FiO2 ratio, another consequence of LRM was that the Crs increment allowed a reduction of ventilation pressure. The importance of this measure during open lung ventilation has already been underscored by Amato et al.,1,2 who established a value of <20cmH2O as one of the main objectives in their ventilation strategy. Conceptually, the relationship between plateau pressure, PEEP and tissue distension can be represented by a simple lever model in which PEEP represents the point of support or fulcrum, plateau pressure is the applied force, the degree of tissue stretch is the displacement reached, and ventilation pressure is the length of the lever torque arm.10 According to this model, on increasing the PEEP level without increasing plateau pressure—as occurred in our patients—LRM decreases the ventilation pressure and consequently also the length of the lever arm and tissue stretch, thereby exerting a favorable influence upon the course of lung damage.2,3

Although the effects and depth of the hemodynamic deterioration of LRM appear to be related to a series of factors, including the basal cardiovascular situation,28–30 the origin and time elapsed from lung injury,17 the type of maneuver employed31 and its duration,29 etc., the marked but generally transient drop in CO is the hemodynamic effect most often reported in different clinical17,32,33 and experimental studies.16,21,28,29,31 In contrast, however, some authors have observed no significant hemodynamic alterations.22,34 On the other hand, the cardiovascular consequences of LRM fundamentally have been attributed to a decrease in venous return and cardiac preloadm,28,32 as well as to an increase in right ventricle afterload22,28—the relative importance of each of these mechanisms being lesser or greater depending on the predominant effect of LRM upon transpulmonary or pleural pressure.35

The results of our study are consistent with the hemodynamic alterations described, with a decrease in CO (of up to 58% in some cases) and left ventricle preload, as assessed from the monitorization of LVETc. However, since right ventricle function was not evaluated, we cannot be sure whether these alterations were fundamentally due to changes in preload or in postload of the right ventricle. Nevertheless, the unequal behavior of Crs over both LRM branches at a comparable PEEP level and mean pressure suggests that transpulmonary pressure was also different in each stage of the maneuver, and therefore that the effect upon right ventricle afterload could be the most relevant factor over the ascending branch, while the decrease in venous return and right ventricle preload as a result of increased transmission to the pleural space could be the most important mechanism over the descending branch of LRM.

In accordance with the observations of previous studies, our results also confirm the scant usefulness of blood pressure as an exclusive parameter for monitoring the hemodynamic consequences of LRM.16,22,31,33,34 In effect, during the end stage of the ascending branch, MBP increased significantly, behaving in a way opposite to CO, and therefore without reflecting the global effect of LRM upon the cardiocirculatory system. Moreover, and curiously, despite the increase in MBP, cardiac power (a parameter assessing the efficiency of the heart as a hydraulic pump, combining simultaneous measurement of flow and pressure)36 decreased significantly over the ascending branch. However, since this parameter appears to be influenced not only by contractility but also by the changes in preload and heart rate,36 we do not know the precise significance of its evolution during LRM, since there were marked changes in these latter parameters during maneuvering. Nevertheless, since acceleration of aortic flow and heart rate increased during the maneuver, we suspect that the changes in cardiac power could be related mainly to worsened preload, and therefore that heart function of the patients was preserved.

On the other hand, although blood gas measurements were not obtained in each PEEP step or in all patients, it can be assumed that progressive hypoventilation with hypercapnia was an accompanying phenomenon during LRM that could have influenced the observed hemodynamic effects. Although there is some controversy regarding the principal hemodynamic effects of acute hypercapnia, they appear to comprise the generation of a hyperdynamic state (high CO, elevated heart rate and lessened systemic vascular resistances), as well as an increase in pulmonary artery pressure that could negatively affect right ventricle function.37–39 It is therefore very probable that all these mechanisms were implicated in the end hemodynamic effect of LRM, and that the combined result of these factors contributed in some way and degree to the observed cardiovascular consequences.

Before drawing conclusions, the present study has important limitations that must be commented. Firstly, this is a retrospective analysis with a limited number of patients—the main cause of respiratory failure being pulmonary infection. It is well known that lung consolidation predominates over collapse in pneumonia. As a result, the recruitment potential is lower than in other forms of respiratory failure, particularly cases of extrapulmonary origin.40 In addition, in the opinion of some investigators, the hemodynamic effects of LRM could be more serious in cases of lung injury induced by an infectious process.31 Lastly, although the hemodynamic monitorization used in this study, in contrast to the intravascular pressures, offers a certain advantage in that it does not interfere with the changes in intrathoracic pressure,19 the behavior of the right-side heart chambers (which are those most affected during LRM, and therefore the main causes of the hemodynamic alterations) could not be evaluated in this study. Despite these limitations, we consider that the study offers relevant information for clinicians wishing to use this form of LRM. Firstly, the maneuver was useful not only for improving oxygenation but also for facilitating selection of open-lung PEEP and lowering lung ventilation pressure. Secondly, since the hemodynamic consequences, and particularly the decrease in CO, were significant and in some cases notorious, we consider that it is not enough to monitor blood pressure during LRM. In this context, careful hemodynamic monitorization should be ensured, if possible using devices that do not generate artifacts with changes in intrathoracic pressure, and which moreover inform us of the course of right ventricle function.

In conclusion, LRM with stepwise increments and decrements in PEEP facilitated open-lung PEEP on the basis of maximum Crs, improving oxygenation and reducing lung ventilation pressure—the main hemodynamic consequence being a reduction in left ventricle preload and cardiac output. The unequal hemodynamic derangement in both LRM branches at comparable PEEP levels and mean airway pressure values in addition showed that apart from intrathoracic pressure, other factors such as Crs and probably hypoventilation with hypercapnia influenced the hemodynamic consequences of this type of maneuver. Lastly, since these alterations were intense and were not reflected in the mean blood pressure changes, it is necessary to emphasize the importance of adequate hemodynamic monitorization, beyond the simple monitoring of blood pressure.

M. Ignacio Monge-Garcia is a consultant to Edwards Lifesciences. The rest of the authors declare no conflicts of interest.