INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Injury caused by or associated with mechanical ventilation (MV) in acute respiratory distress syndrome (ARDS) has become a subject of great concern (1). Various experimental studies have described the physiologic mechanism by which MV may lead to ventilation-induced lung injury (VILI) (2, 3). The factors that mainly contribute to VILI include high distending transalveolar pressure or overdistension, related to high pressures and volumes that occur at the end of inspiration (4), and transalveolar pressure falling below the critical closing pressure of alveolar units at the end of expiration, inducing repetitive opening-closing phenomenon (3, 5, 6). Although these factors occur either at high lung volume (i.e., overdistension) or at low lung volume (i.e., opening closing), they are often associated throughout the respiratory cycle in the ARDS lung because of the uneven distribution of lung disease.

On the basis of these clinical and experimental factors, it is now recommended that, in order to limit VILI, plateau pressure should be kept below 30 to 35 cm H₂O while maintaining the lung open with sufficient positive end-expiratory pressure (PEEP) (7). Reduction of tidal volume (VT) is the cornerstone of this strategy, and PEEP should be set at a sufficient level in order to avoid end-expiratory collapse.

Three recent controlled trials that compared low (6 to 8 ml/kg) versus conventional (10 to 11 ml/kg) VT for a given PEEP level set empirically, failed to demonstrate any benefit of a pressure-limited strategy with regard to morbidity and mortality (8). In contrast, two other studies showed a significant benefit on mortality by reducing VT (11, 12). One of these studies associated the reduction in VT to a high PEEP level, whereas the most recent and largest trial to date has used a high respiratory rate in the low VT group, which could lead to gas trapping and a higher total PEEP. In addition, a recent clinical trial demonstrated that the use of high PEEP and low VT reduced inflammatory cytokines level both at the bronchoalveolar lavage and in the blood (2). As previously suggested, a high PEEP level could be particularly important in this low VT strategy, because hypoventilation may lead to progressive derecruitment. For this reason, it has been proposed to add recruitment maneuvers to the standard approach of reduced VT (12).

The aim of this study was to test the hypothesis that, for a total PEEP kept constant at the lower inflection point (Plip) on the pressure-volume (Pel-V) curve of the respiratory system, a reduction in VT from 10 to 6 ml/kg could be responsible for alveolar derecruitment. The respective efficacy of a recruitment maneuver and an increase in PEEP level above Plip was also assessed by means of the Pel-V curve.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Patients requiring MV for more than 24 h and fulfilling the criteria for acute lung injury as defined by a Pa_O2/FI_O2 ratio < 300 mm Hg, bilateral opacities on chest radiograph, and no history suggesting elevated left atrial pressure, were candidates for inclusion (13). Patients presenting with a documented history of chronic obstructive pulmonary disease or a contraindication for sedation and paralysis were not included in the study. The protocol was approved by the Henri Mondor Hospital Ethics Committee, and informed consent was obtained from patient's next of kin.

Fifteen consecutive patients were enrolled in the study. All patients were sedated, paralyzed, and ventilated in the volume-controlled mode (Servo Ventilator 900C; Siemens-Elema AB, Solna, Sweden).

Pel-V Curve Recording

The system, including a computer-controlled Servo Ventilator 900C, and the technique for performing Pel-V curves, based on the low-flow insufflation method, have been previously described in detail (14). This allowed Pel-V curves to be obtained either from PEEP or from zero end-expiratory pressure (ZEEP).

Each Pel-V curve was analyzed according to a mathematical model that divided the curve into three segments, separated by a lower and upper inflection point (LIP and UIP, respectively) (15). (See online data supplement.)

Alveolar Recruitment

PEEP Pel-V curve was plotted on the same volume axis as the ZEEP Pel-V curve, using PEEP-related end-expiratory lung-volume variation measured during the passive expiration from PEEP to ZEEP. PEEP-related recruitment was defined, for a given elastic pressure (Pel), by the volume difference between both curves, taking into account intrinsic PEEP (PEEPi) (14). This volume represented the PEEP-related recruitment of previously collapsed lung units, and was identified by the upward shift along the volume axis of the PEEP Pel-V curve, relative to the ZEEP Pel-V curve. Recruitment could be quantified at any elastic pressure studied.

Protocol

The first Pel-V curve was obtained from ZEEP, without modifying previous ventilator settings, in order to determine pressure at the LIP (Plip). The total PEEP (PEEPtot) level (taking into account PEEPi) was set close to Plip and kept constant whatever the VT tested, by modifying external PEEP level when necessary. The respiratory rate (from 18 to 22 min¹), the inspiratory/total time ratio (0.33), and the FI_O2 were also kept constant.

First part of the study (15 patients). Patients were successively ventilated with conventional VT (CVT = 10 ml/kg of "dry" body weight) and low VT (LVT = 6 ml/kg) randomly applied. Patients were ventilated 1 h in each mode. After sampling blood for arterial blood gas measurements, the following measurements were made during the last 10 min of each condition (CVT and LVT). (1) The PEEP Pel-V curve was recorded after a prolonged expiration during which PEEP was maintained. (2) The ZEEP Pel-V curve was then recorded after a prolonged expiration during which PEEP was eliminated (i.e., from the elastic equilibrium volume reached at ZEEP).

Second part of the study (10 patients). Similar measurements were made, respectively, with CVT and LVT immediately after two successive recruitment maneuvers consisting of a 45 cm H₂O pressure-limited breath applied for 15 s.

Third part of the study (10 patients). The PEEP level was then increased by 4 cm H₂O, taking into account PEEPi to keep PEEPtot constant whatever the VT. Pel-V curve recordings and recruitment calculation were repeated after 1 h of CVT and LVT ventilation combined with the high level of PEEP.

Statistics

Values are given as means ± SD. At a given PEEP level, alveolar recruitment (calculated for any given Pel), ventilatory settings, hemodynamic data, and arterial blood gases measured with CVT and LVT, were compared using Wilcoxon's test for paired samples. The effect of a recruitment maneuver and increasing PEEP was compared with those observed before recruitment maneuver or PEEP increase using the same test. Significant differences were considered for p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Arterial blood gases and Pel-V curves recorded from PEEP set at Plip, and from ZEEP, were obtained in 15 patients with CVT and LVT, respectively. The mean LIP was 10 ± 4 cm H₂O. Pel-V curves were also recorded after a recruitment maneuver and after an increase in PEEP by 4 cm H₂O in 10 patients of the whole population. The characteristics of the patients are presented in Table 1.

Influence of VT Keeping PEEP Constant

Both strategies were well tolerated by all patients. Ventilatory settings, plateau pressure (Pplat), setPEEP, and PEEPtot are given in Table 2. PEEPtot observed with CVT did not differ from LVT, and this was achieved by individual adaptation of external PEEP level. As expected, Pa_CO2 was significantly higher with LVT (Table 3). Hemodynamic parameters were not affected by change in VT (Table 3).

Over the Pel range from 15 to 30 cm H₂O, recruitment was significantly higher with CVT than with LVT, for a similar PEEPtot level. Recruitment calculated for 20, 25, and 30 cm H₂O of Pel, is illustrated in Figure 1. Expressed as a percentage of the amount of recruitment observed with CVT, recruitment with LVT was only 69% at 15 cm H₂O and 59% at 30 cm H₂O. Pa_O2, as well as Pa_O2/FI_O2, were not significantly altered by reduction of VT. SaO₂ was significantly lower with LVT than with CVT (Table 3).

View larger version (28K): [in this window] [in a new window]   Figure 1.   Alveolar recruitment, measured as the volume difference between pressure-volume curves traced from PEEP and from ZEEP, expressed over the range of distending pressures (Pel) where both curves are traced during low tidal volume (LVT) and conventional tidal volume (CVT), respectively, and using the same PEEP level. Values are expressed as means and standard deviations. On the left, the baseline PEEP level set at the lower inflection point (LIP) is used; in the middle, the same measurements are performed immediately after a recruitment maneuver; on the right, PEEP has been increased by 4 cm H2O. Significant differences between CVT and LVT were found only in the first situation. *Indicates significant difference (p < 0.05) between LVT and CVT; **p < 0.01.

Effect of a Recruitment Maneuver on the Derecruitment Induced by Low VT

In the 10 patients in whom a recruitment maneuver was performed, PEEPtot set at LIP and Pplat were: 13 ± 3 versus 13 ± 4 cm H₂O, NS, and 36 ± 7 versus 28 ± 7 cm H₂O, p < 0.01, for CVT and LVT, respectively. The comparison of recruitment between CVT and LVT before the recruitment maneuver showed the same difference as in the 15 patients.

With LVT, a recruitment maneuver induced a significant increase in recruitment (175 ± 108 ml versus 254 ± 137 ml, expressed at a Pel of 20 cm H₂O, p < 0.01). In contrast, the same recruitment maneuver did not significantly affect recruitment with CVT (266 ± 157 ml versus 264 ± 120 ml, NS). As a result, PEEP-induced recruitment became similar for LVT and CVT after the recruitment maneuver (Figure 1).

Effect of Increasing PEEP on the Derecruitment Induced by Low VT

In the same 10 patients, the 4 cm H₂O increase in PEEP modified PEEPtot and Pplat to 17 ± 3 versus 16 ± 3 cm H₂O, NS, and to 41 ± 9 versus 32 ± 7 cm H₂O, p < 0.01 for CVT and LVT, respectively.

The increase in PEEP induced an increase in recruitment for both VT (expressed at 20 cm H₂O of Pel: 175 ± 108 versus 332 ± 91 ml with LVT, p < 0.05, and 266 ± 157 versus 464 ± 216 ml with CVT, p < 0.05) (Figure 1). At this high PEEP level, the difference in recruitment between LVT and CVT (Figure 1) did not reach significance.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study has demonstrated that a reduction in VT from 10 to 6 ml/kg, keeping PEEPtot constant at the LIP, was responsible for a significant lung volume loss corresponding to alveolar derecruitment. This effect may have participated, at least in part, to the negative results of three recent clinical trials that failed to demonstrate any benefit of a ventilatory strategy aiming at systematically reducing VT (8). This hypothesis is corroborated by the significant reduction in mortality rate associated with a lung protective strategy recently reported by Amato and colleagues (12). In this study, particular attention was given to the PEEP setting, as well as to repeated sighs, performed to preserve lung recruitment during low VT ventilation. In contrast, the recent NIH Network trial seems to indicate that reduction in VT alone is effective to reduce mortality. However, the possibility of a PEEPi effect in the low VT group, caused by the high respiratory frequencies used, makes the interpretation of these results still open.

The apparent conflicting results of these recent clinical trials could also suggest that, first, reduction in VT from 10 to 6 ml/kg alone fails to improve mortality and, second, that a specific strategy to avoid harmful alveolar derecruitment may be beneficial. In addition, a recent clinical study has indicated that a low PEEP associated with high VT may lead to systemic inflammatory response and possibly promotes lung injury (2).

Alveolar Derecruitment

Several studies based on CT scan, Pel-V curves, or gas exchange have demonstrated that recruitment is a continuous and progressive phenomenon that not only depends on PEEP but also on peak inflation volume (14, 16). Considering a given PEEP level, if tidal inflation induces recruitment, a reduction in VT could thus be responsible for alveolar derecruitment. It has been demonstrated that low VT ventilation during anesthesia could be responsible for alveolar collapse in healthy lungs (19, 20). Several factors, including the use of high FI_O2 in the presence of low / ratio, the use of low VT, sedation, and paralysis may facilitate atelectasis and promote denitrogenation atelectasis in patients with ARDS (21). Cereda and colleagues (24) have shown that low VT ventilation could induce a progressive decrease in compliance, indicating a time-dependent derecruitment, which could be prevented by higher PEEP level. The influence of VT on PEEP-related cardiorespiratory effect was also studied by Ranieri and colleagues (25) in a group of nine patients with ARDS. Recruitment based on Pel-V curve analysis was measured in each patient during low and high tidal ventilation for a constant PEEP level. In all patients, the Pel-V curve from ZEEP during low VT ventilation exhibited a concavity toward the volume axis, which suggested a recruitable lung, whereas the Pel-V curves recorded with high VT were convex in six patients, indicating a progressive decrease in compliance suggesting overdistension. These investigators concluded that with high VT, PEEP mainly induced hyperinflation of alveoli already recruited by tidal ventilation, whereas with low VT, PEEP induced alveolar recruitment and counterbalanced low VT-related derecruitment. Despite some discrepancies with our results, possibly related to methodologic differences, the results reported by Ranieri and colleagues showed that low VT ventilation might affect lung recruitment. A recent study corroborating these findings suggested that the size of VT might interfere with recruitment by modifying the behavior of airway resistance and time constant (26). In the present study, the significant alveolar derecruitment associated with VT reduction was recovered by increasing PEEP, which confirmed that the lung was not fully recruited during low VT ventilation. However, because we kept VT constant at the two different PEEP levels (either with LVT or with CVT), the Pplat increased with the higher PEEP level. Therefore, we cannot determine whether the greater recruitment was a result of the Pplat causing inflation of more lung units at end-inspiration or higher PEEP preventing more time dependent end-expiratory collapse or both.

Influence on Oxygenation

It is known that oxygenation depends on recruitment, and therefore a worsening in oxygenation could have been expected with low VT. In the present study, Pa_O2 did not significantly change with the reduction in VT but arterial oxygen saturation worsened. A right-shift in the oxyhemoglobin dissociation curve caused by acidosis may have participated to this latter finding. However, the influence of VT reduction on gas exchange remains controversial in the studies that have specifically addressed this issue. Several mechanisms, not directly related to recruitment, could be involved in the oxygenation changes occurring after VT reduction. Hypercapnia associated with VT reduction tends to increase cardiac output (27). This effect may improve Pa_O2 by increasing mixed venous oxygen saturation or decrease Pa_O2 by increasing shunt. In the study reported by Ranieri and colleagues (26), Pa_O2 was significantly improved when VT was reduced from 12 to 7 ml/kg. These results greatly differ from those of Blanch and colleagues (30) who reported a better recruitment as well as a Pa_O2 improvement associated with an increase in VT. In this study, minute ventilation was matched by increasing respiratory rate with low VT in order to limit hypercapnia induced by VT reduction. Kiiski and colleagues (31, 32) studied the influence of VT on gas exchange and oxygen delivery during, respectively, nonmatched and matched minute ventilation. Both studies suggested that changes in gas exchange observed with low VT not only depend on recruitment but also on the hemodynamic effects associated with hypercapnia. Our results concerning oxygenation should be interpreted with caution because of the relatively small number of patients. The absence of invasive hemodynamic monitoring makes it difficult to interpret the effects of both ventilatory strategies on oxygenation.

Recruitment Maneuver

The lung protective approach reported by Amato and colleagues (12) included the application of systematic sighs after tracheal suctioning, with the aim of avoiding harmful alveolar collapse. The rationale for using a recruitment maneuver is to forestall collapsed lung units, which suggests that this technique is effective when the lung has previously been derecruited. Several studies have demonstrated the effect of sighs on atelectasis observed during general anesthesia (19, 20). Pelosi and colleagues (33) investigated the effects of sighs on gas exchange in patients with acute lung injury. These investigators demonstrated that application of sighs during lung protective strategy could improve lung volume and oxygenation. These results indicate that a low VT strategy may not provide full lung recruitment despite the relatively high PEEP level (14 ± 2 cm H₂O) applied. A recent study confirmed these findings but also suggested that recruitment maneuvers were less efficient than PEEP increase (34). In our study, the increase in recruitment resulting from the recruitment maneuver was observed only during low VT ventilation. Interestingly, the same maneuver did not lead to any significant beneficial effect during CVT ventilation, indicating that the lung remained open and therefore was no longer recruitable. However, time effect of recruitment maneuver could not be evaluated since ventilatory settings were changed to higher PEEP level immediately after measurements.

In conclusion, the most relevant finding of this study was that a ventilatory strategy based on recent recommendations (i.e., limited VT and PEEP at or above Plip) leads to alveolar instability and lung collapse. These results suggest that a specific approach may be needed to prevent alveolar instability in the situation where strategies limiting end-inspiratory pressure and overdistension are used. In our study, increasing PEEP or performing a recruitment maneuver appear as two possible strategies to counteract low VT-induced derecruitment.