INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the setting of unilateral lung injury, measurement of global respiratory system mechanics does not provide clinically useful information to set ventilator parameters (PEEP or tidal volume) because the mechanical impairment of the injured parts of the lung cannot be specifically assessed (1, 2). In fact, some patients with acute respiratory distress syndrome (ARDS) presenting with consolidated lower lung lobes that cannot be recruited present a gas exchange response to increasing levels of positive end-expiratory pressure (PEEP) similar to that described in patients with unilateral lung injury (3). In this context, application of PEEP usually does not improve oxygenation as it may cause overdistension of compliant lung regions and redistribute blood flow to collapsed or fluid-filled alveoli (4).

Increasing evidence has shown that both cyclical closing and reopening of alveolar units and elevated alveolar pressures associated with overdistension of lung units can cause ventilator-associated lung injury (VALI) (8). It is difficult to adjust tidal volume (VT) and PEEP to avoid VALI in unilateral lung injury. Preferential distribution of ventilation promote overdistension of the normal lung, whereas attempts to minimize it by decreasing PEEP may then lead to repetitive closure and opening of lung units in the injured lung. Selective application of tracheal gas insufflation (TGI) can potentially eliminate this problem by creating selective autoPEEP only in the injured lung.

TGI has proved to be a useful adjunct to mechanical ventilation in patients with acute respiratory failure. TGI can be used either to decrease Pa_CO2 in the setting of hypercapnia or to limit ventilatory distending forces by allowing a reduction in VT while maintaining constant Pa_CO2 (11). In severe unilateral lung injury, we hypothesized that selective application of TGI to the diseased lung (1) may increase lung volume of collapsed regions without adversely affecting blood flow or causing regional lung overdistension and (2) may allow a reduction in VT while maintaining eucapnia, thus decreasing the potential for VALI. Consequently, the objective of our study was to assess the effects of selective pan-expiratory TGI on lung function in a model of unilateral lung injury induced by saline lavage of the left lung in dogs.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Preparation

We studied eight supine mixed-breed dogs weighing 27.8 ± 4.3 kg (mean ± SD) using a protocol approved by the Animal Care and Use Committee at Regions Hospital. The dogs were mechanically ventilated (Veolar; Hamilton Co., Reno, NV) in the constant-flow, volume-cycled mode with a VT of 15 ml/kg and a frequency of 14 breaths/ min. We used a PEEP of 3 cm H₂O and an inspired O₂ fraction (FI_O2) of 0.80 to maintain adequate oxygenation. Subsequently, the dogs were tracheostomized, reintubated through the tracheotomy by a double-lumen endotracheal tube to isolate the two lungs, and connected to a single ventilator (Veolar; Hamilton) by a double inverted Y. Once saline lavage of the left lung was completed, the double- lumen endotracheal tube was removed and the animals were reintubated by a single lumen endotracheal tube (Univent^®; Vitaid, Lewiston, NY) through the tracheotomy. The Univent single lumen tube incorporates within its main channel an internal catheter (ID, 2.0 mm) that can be extended 10 cm beyond the distal tip of the endotracheal tube (Figure E1, in the online data supplement) (16, 17). The tip of the small catheter within the Univent tube was extended and positioned 1 to 2 cm below the main carina in the left main bronchus under bronchoscopic guidance. We used the internal catheter for selective delivery of TGI flow to the injured left lung.

Experimental Protocol

Prior to initiation of lung injury, measurements of hemodynamics, gas exchange, and lung and chest wall mechanics, including a FRC at ZEEP and a pressure-volume (P-V) curve were performed (18). Subsequently, we injured the left lung as described above. After stabilization (less than a 5% change in the measured Pa_O2 over a 15-min period), a chest radiograph was performed to confirm the unilaterality of the lung injury. Once the injury was stable, we obtained a set of measurements (hemodynamics, gas exchange, P-V curve, and lung and chest wall mechanics) at a PEEP of 3 cm H₂O. Afterwards, PEEP was increased from 3 to 10 cm H₂O and measurements were repeated after a stabilization period of 20 min. We then applied expiratory TGI to the left lung and progressively increased catheter flow rate to a maximum flow of 25 L/min. Because the increment in end-expiratory lung volume (EELV) induced by TGI was continuously assessed by inductive plethysmography, we could maintain EELV constant throughout the experimental protocol by decreasing the PEEP set on the ventilator. Measurements were repeated after a stabilization period of 20 min. To assess the ability of selective TGI to decrease airway pressures we decreased VT, progressively targeting a Pa_CO2 value obtained at baseline (at PEEP 10 cm H₂O without TGI). During this stage EELV was maintained constant by adjusting ventilator-set PEEP if necessary. Finally, TGI was turned off, PEEP was set at 3 cm H₂O, and VT was increased to 15 ml/kg. In this manner, all the TGI stages were bracketed by a conventional mechanical ventilation stage to ensure the stability of the experimental model.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of Unilateral Lavage

Left lung lavage significantly reduced Pa_O2/FI_O2 from 483 ± 5 to 164 ± 36 mm Hg (p < 0.01), and increased intrapulmonary shunt from 9 ± 1 to 43 ± 6% (p < 0.01). FRC decreased from 690 ± 40 to 451 ± 52 ml (p < 0.05) and the presence of pulmonary infiltrates, confined solely to the left lung, and loss of lung volume were radiologically confirmed after left lung lavage. After unilateral lung injury, airway pressures increased and total respiratory system and lung compliance decreased. Chest wall compliance remained essentially the same. Mean Ppa increased after unilateral injury, whereas the rest of the hemodynamic parameters did not change. Hemodynamic and respiratory mechanics data before and after unilateral injury are depicted in Tables 1 and 2. Lower and upper inflection points measured from the static P-V curves constructed in seven animals are shown in Table 3. Two representative P-V curves done in basal conditions and after left lung injury are depicted in Figure 1.

View larger version (28K): [in this window] [in a new window]   Figure 1.   Pressure-volume curve of the respiratory system obtained in two representative animals before (open circles) and after (crosses) acute left lung injury (lung lavage) showing two inflection points (lower and upper).

Effect of Increasing PEEP

Changing PEEP from 3 to 10 cm H₂O significantly increased airway pressures but did not alter hemodynamics and failed to improve oxygenation or intrapulmonary shunt (Figure 2). Alveolar ventilation and physiologic dead space did not change (Figure 3). The effects of increasing PEEP on hemodynamics and respiratory system mechanics are depicted in Tables 1 and 2. Lung and chest wall compliance at PEEP 10 cm H₂O were not different compared with PEEP 3 cm H₂O.

View larger version (35K): [in this window] [in a new window]   Figure 2.   Oxygenation (PaO2/FIO2) and calculated venous admixture ( VA/ T) at baseline (BSL), after lung injury (lavage) at a PEEP of 3 cm H2O and 10 cm H2O (PEEP 3 and PEEP 10, respectively), during expiratory tracheal gas insufflation to the left lung (TGI), during expiratory tracheal gas insufflation to the left lung associated with lower tidal volume (TGI+Low VT), and during PEEP of 3 cm H2O (PEEP bracket). (a) Significant difference between baseline values compared with the rest of the stages. (b) Significant difference between TGI and TGI+Low VT values compared with the rest of the stages.

View larger version (36K): [in this window] [in a new window]   Figure 3.   PaCO2 and calculated physiologic dead space (VD/VT) at baseline (BSL), after lung injury (lavage) at PEEP = 3 cm H2O (PEEP 3), during expiratory tracheal gas insufflation to the left lung (TGI), during expiratory tracheal gas insufflation to the left lung associated with lower tidal volume (TGI+Low VT), and during PEEP = 3 cm H2O (PEEP 3 bracket). (a) Significant difference compared with baseline. (b) Significant difference compared with TGI. N/A = not available (could not be calculated accurately because of small tidal volumes).

Effect of Selective TGI Application

During selective TGI, ventilator set PEEP was decreased to 4.9 ± 0.5 cm H₂O to maintain EELV at the same level as that obtained during ventilation at a PEEP of 10 cm H₂O without TGI. Selective left lung TGI significantly improved Pa_O2/FI_O2 from 212 ± 43 mm Hg at a PEEP of 10 cm H₂O, to 301 ± 38 mm Hg (p < 0.01), and decreased venous admixture (VA/T) from 33 ± 5 to 20 ± 2% (p < 0.05) (Figure 2) without any change in hemodynamics (Table 1). Pa_CO2 and VD/VT decreased significantly (p < 0.01) compared with PEEP = 10 cm H₂O (Figure 3). Airway pressures (Ppeak and Pplat) markedly decreased during selective TGI compared with PEEP = 10 cm H₂O (p < 0.01) (Table 2 and Figure E2 in the online data supplement). AutoPEEP increased from 0.2 ± 0.1 cm H₂O at PEEP = 10 cm H₂O to 3.8 ± 0.6 cm H₂O during selective TGI (p < 0.01). Total PEEP during selective TGI was slightly lower, although it was not statistically different when compared with PEEP = 10 cm H₂O. During TGI, chest wall compliance did not change and lung compliance increased by 45% but did not reach statistical significance (Table 2).

Effect of Decreasing VT During Selective TGI

During ventilation with selective TGI, tidal volume was reduced to 5.2 ml/kg (145 ± 34 ml) without changing EELV. AutoPEEP was 3.5 ± 0.5 cm H₂O. Pa_O2/FI_O2 and VA/T remained unaltered (Figure 2), whereas Pa_CO2 increased to the same level as that observed during the previous stages (PEEP 10 cm H₂O and selective TGI at normal VT) (p < 0.05) (Figure 3). At this stage, airway pressures decreased significantly (Figure E2) and were not different from the baseline airway pressures measured during pre-injury stage (Table 2). Lung compliance significantly improved compared with PEEP = 3 cm H₂O and PEEP = 10 cm H₂O stages (p < 0.01) (Table 2). Hemodynamic parameters remained essentially unaltered.

Oxygenation, alveolar ventilation, respiratory system mechanics and hemodynamics returned to those observed at previous baseline stage (PEEP = 3 cm H₂O) when the baseline stage (PEEP 3 Bracket) was repeated after the TGI stages (Tables 1 and 2, and Figures 2 and 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has shown that in unilateral lung injury, selective TGI improved oxygenation as compared with conventional mechanical ventilation with global PEEP. The improvement in oxygenation was accomplished at a lower global PEEP, but at the same total EELV. This suggests that regional autoPEEP provided by selective TGI recruited the injured lung, thereby protecting the normal lung from the adverse effects of lung overdistension. Additionally, selective expiratory TGI reduced Pa_CO2 by 30% while maintaining minute ventilation constant, and at similar levels of alveolar ventilation allowed a 60% reduction in tidal volume. Consequently, in unilateral lung injury, selective TGI can be used either to limit the extent of hypercapnia or the magnitude of tidal volume.

Selecting PEEP in Patients with Unilateral Lung Injury

Mechanical ventilation in the context of severe unilateral lung injury is often unable to reverse the severe hypoxemia caused by increased shunt through the affected lung. Ventilation with PEEP may cause distension of the normal lung, whereas the injured less compliant lung will be much less expanded (5). This may result in redistribution of blood flow from the overdistended normal lung to the injured lung, thereby increasing the degree of intrapulmonary shunt. Currently, differential ventilation with selective PEEP (1, 23) and positioning the patient with the "good lung down" (6, 24) are the two accepted treatments to improve oxygenation in severe unilateral lung injury. However, both techniques have serious limitations in clinical practice. Independent lung ventilation with selective PEEP may cause tracheobronchial damage, requires muscle relaxation and two ventilators, and poses difficulties in suctioning the airways. Positioning the patient with the "good lung down" has the risk of soiling the good lung with worsening overall lung function and cannot be sustained for long periods.

Measurement of respiratory system mechanics in order to optimize mechanical ventilation in patients with acute lung injury has been proposed to improve oxygenation and avoid VALI (9). These goals may be accomplished by adjusting the ventilator settings according to the P-V curve of the respiratory system. This approach, although useful in patients with ARDS (10), provides much less information in unilateral or localized lung injury. In patients with severe unilateral pneumonia, Carlon and colleagues (1) demonstrated that a high global PEEP level caused distension of the good lung without allowing ventilation of the diseased one. This effect was noticed only when P-V curves of the right and left lungs were performed separately. The application of an optimal PEEP to each lung yielded similar unilateral P-V curves, indicating improvement of the mechanics of the diseased lung without overdistending the normal lung. Similarly, Vieira and colleagues (7), studying patients with ARDS with P-V curves that lacked a lower inflection point (LIP), indicating the absence of recruitable volume, demonstrated that the application of PEEP overdistended aerated lung regions without opening consolidated areas, and improvement in oxygenation was not observed.

We performed supersyringe P-V curves (Figure 1) in our experimental animals with unilateral lung injury (lung lavage) and found a LIP of 20.2 ± 1.3 cm H₂O (Table 1). LIP measured during unilateral injury was higher than the LIP observed in dogs with bilateral lung lavage injury (25), and suggests that LIP signals the beginning of alveolar recruitment of the compliant regions of the diseased lung (26, 27). Interestingly, this occurred at relatively high pressures in unilateral lung lavage because in unilateral injury the normal lung must be inflated to near total lung capacity before recruitment of the injured lung can take place. The presence of an upper inflection point indicates the end of the recruitment process and the beginning of global overdistension (26, 27). Accordingly, global P-V curves in unilateral lung injury are influenced by the marked heterogeneity of the disease process and cannot be used to choose an appropriate PEEP level.

Application of Selective TGI

The overall goal of TGI is to reduce distending forces acting on the lung while improving gas exchange. Expiratory TGI improves the efficiency of alveolar ventilation by facilitating the clearance of CO₂ from anatomic dead space, and increases EELV in a flow-dependent fashion (11, 22). Catheter flow can increase alveolar pressure in three ways (28). First, part of the momentum of the discharging jet stream is transferred to the alveoli. Second, placement of the catheter decreases the cross-sectional area of the trachea available for expiratory flow, effectively increasing expiratory resistance. Third, catheter flow through the endotracheal tube, expiratory circuit, and expiratory valve during expiration builds a back pressure at the airway opening that impedes expiratory flow from the lung. If TGI is applied selectively by insufflating the gas directly to one lung these effects will be confined primarily to that lung.

In our study during selective TGI application, overall EELV was kept constant by decreasing ventilator set PEEP. Interestingly, both gas exchange and global lung compliance were better at lower airway pressures with TGI as compared with ventilation at PEEP = 10 cm H₂O. These findings suggest that distribution of volume within the lung was more homogeneous during selective TGI application. In other words, the improvement in lung compliance during TGI was due to avoidance of overdistension of the normal lung and recruitment of the diseased lung, distributing the VT more homogeneously between the two lungs. The fact that hemodynamics was not impaired and VA/T was markedly reduced indicates better matching of ventilation and perfusion during selective TGI in the setting of unilateral lung injury.

During selective TGI, autoPEEP increased significantly from 0.2 ± 0.1 to 3.6 ± 0.6 cm H₂O, suggesting that expiratory flow from the injured left lung was impeded by TGI jet discharging within the lumen of the main left bronchus. Local autoPEEP within that lung increased left lung volume at end-expiration and moved that lung to the more compliant part of its P-V curve, allowing better tidal ventilation with each inspiration (29). Consequently, better matching of left and right lung volumes was accomplished during selective TGI compared with conventional mechanical ventilation at the same global EELV.

In patients with ARDS, part of the dead space resides in the alveoli as alveolar dead space (30, 31). The alveolar gas originating from those ventilated but hypoperfused lung regions were CO₂-poor, diminishing the impact of washing proximal dead space free of CO₂ on alveolar ventilation (28, 32). The effect of positioning the catheter tip 1 to 2 cm below the main carina on Pa_CO2 is minimal under normal conditions (12). Despite having the left lung exhaling relatively CO₂-poor gas, the effects of the TGI were extended to the trachea, washing out the CO₂-laden gas residing in the proximal anatomic dead space exhaled primarily from the normal lung. This effect most likely does not fully explain the marked reduction in Pa_CO2 and VD/VT observed during selective TGI despite moderate levels of hypercapnia in our experimental model. Part of the CO₂ clearance efficacy of TGI was probably related to the ability of regional PEEP created by selective TGI to convert shunt units that were not exhaling any CO₂ into open units participating in gas exchange, thus increasing CO₂ clearance.

Application of Selective TGI at Low Tidal Volume

Overdistension of lung units and repetitive alveolar collapse and reopening are thought to be the factors responsible for the development of VALI (8). However, VALI might also occur in patients with acute unilateral lung injury. Alveolar overdistension and repetitive recruitment and derecruitment are not limited to certain anatomic locations, i.e., nondependent and dependent lung regions, respectively. In unilateral lung injury, alveolar overdistension can occur in the healthy lung with simultaneous alveolar collapse and reopening in the poorly recruited diseased lung. Ultimately, the magnitude of this phenomenon will depend on the distribution of EELV and VT between the two lungs.

Several studies (15, 33) have shown that one of the most important features of TGI is to maintain normocapnia or a given level of Pa_CO2 while tidal volume is decreased, allowing a reduction in minute ventilation. TGI can thereby be used to decrease the forces acting on the lung thus minimizing VALI in patients. In our study, we found that tidal volume could be reduced from 15 to 5.2 ml/kg during selective TGI application, decreasing airway opening pressures by ~ 60% without affecting gas exchange. Although left lung compliance was maximal during TGI with low VT, we cannot fully exclude overdistension of certain portions of the left lung during the tidal cycle. Conceivably, some pendelluft could have occurred at the beginning of inspiration until the pressures within the two lungs became equal. However, this effect is at best small, as decompression of the TGI circuit and closure of the TGI gating solenoid valve is not instantaneous. Similarly, plateau pressure at end-inspiration reflects an average pressure in the system and not the pressures within each lung.

Clinical Implications

Currently, there is not a standard method of introducing an insufflation catheter into the trachea. In most human studies, a small caliber catheter has been introduced through an angled side-arm adapter attached to the endotracheal tube and positioned above the carina (13, 32), or incorporated into the wall of the endotracheal tube (14, 30, 34). Endotracheal tube designs that incorporate channels within the endotracheal tube wall (14, 30, 34) are not suitable for selective TGI application. In this study, we intubated the experimental animals with an endotracheal tube with a movable bronchial blocker (Univent). This endotracheal tube has a small channel through the anterior internal wall, which contains an endobronchial blocker 2 mm in diameter with a low pressure cuff. The endobronchial blocker can be fully retracted into the main body of the endotracheal tube when not in use, or it may be advanced into the right or left mainstem bronchus blindly or under bronchoscopic control. The Univent tube is widely and safely used (instead of double-lumen tubes) for one-lung anesthesia, and to prevent aspiration in cases of massive hemoptysis (16, 17, 35). In critically ill patients, application of unilateral TGI could be safer than unilateral PEEP applied via a divided airway.

Selective TGI was successfully applied during a short time interval using the Univent tube in our study. Although TGI improved pulmonary function without any side effect, it is hazardous to extrapolate our experimental data to the clinical setting. Potential concerns of selective TGI at high flows include bronchial mucosal damage as well as inspissation or retention of secretions, especially if the insufflated gas is not adequately humidified. Heating and humidifying insufflated gas is possible, but the high pressures that are developed when gas is forced through the small-bore catheters may exceed a humidifier's leak or burst pressure (36). Finally, we do not know whether the high TGI flows used in this study could be maintained for a long period.

Severe unilateral lung pathology such as unilateral pneumonia or refractory atelectasis are common conditions in the intensive care setting, often requiring mechanical ventilation with high FI_O2. Moreover, when unilateral injury coexists with cerebrovascular disorders, uncorrected hypovolemia, or severe metabolic acidosis, it may be clinically desirable to normalize CO₂ while avoiding the use of large inflation volumes to minimize ventilator-associated lung injury (37). In these clinical scenarios, regional recruitment with selective TGI may be a useful clinical tool, although clinical studies are warranted to assess the safety and validity of this novel treatment.