INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

The recent recognition of the potential for ventilator-induced lung injury (VILI) in patients with acute respiratory distress syndrome (ARDS) has focused attention on ventilatory strategies aimed at reducing airway pressure and tidal volumes (VT) (1).

Almost invariably, however, low VT ventilation results in hypercapnia (4), which, in turn, carries many side effects, such as pulmonary hypertension and increased intracranial pressure in patients with head trauma. Reducing the potential for barotrauma and VILI remains, however, of major importance in the management of patients with ARDS.

Tracheal gas insufflation (TGI) improves CO₂ elimination by its effect of bypassing or washing out the anatomical dead space. The addition of TGI techniques to mechanical ventilation allows the use of very small VT that, when combined with an appropriate respiratory rate, may reach the target of avoiding dangerous ventilatory settings while maintaining normocapnia. TGI techniques have been developed and investigated both in animal (5) and human studies (9).

Intratracheal pulmonary ventilation (ITPV) is a TGI technique that delivers the entire VT in the proximity of the tracheal carina, while the ventilator acts as a shutter that directs the flow alternatively in and out of the respiratory system (16).

To our knowledge, the implementation of continuous TGI on patients with acute respiratory failure has been tested as an adjunctive support to mechanical ventilation at relatively low flow rates only (up to 6 L/min) (13, 15, 19). ITPV alone or as an adjunct to mechanical ventilation has been applied on patients with acute lung injury, although clinical experience is still limited (20, 21). A continuous flow of fresh gas as the only source of ventilation has never been evaluated systematically on sedated and paralyzed patients with ARDS. The present study evaluates the CO₂ clearance efficiency of continuous ITPV delivering the entire VT through the intratracheal catheter and compares it with volume-cycled mechanical ventilation.

A possible major problem with ITPV in adult patients is related to the high gas flow that is needed at the carina to provide the entire VT: high flows, when delivered through standard catheters, commonly create an obstacle to exhalation, resulting in hyperinflation and incomplete expiration.

A special intratracheal reverse-thrust catheter (RTC) has been designed to direct the gas flow out of the lung. Such a catheter has been proven to help the expiratory phase, avoiding dynamic hyperinflation through a Venturi effect (17, 22). To verify this additional benefit, we used an RTC in the present study, and in a subgroup of patients we also measured end-expiratory lung volume (EELV) by respiratory inductive plethysmography (RIP).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Eleven patients meeting the ARDS criteria (23) with a Murray score of 2.5 ± 0.3 at the time of the study were enrolled. All patients were under stable conditions without significant variations in hemodynamic status or body temperature. In nine patients a thermodilution Swan-Ganz catheter was in place before enrollment, and all had an intraarterial catheter and a central vascular line for clinical monitoring. Patients' demographics and clinical data are shown in Table 1. No invasive procedures were performed for at least an hour before data collection to limit baseline metabolic rate and CO₂ elimination (CO₂) variations.

The experimental protocol was reviewed and approved by the Ethical Committee of the S. Gerardo Hospital and informed consent was obtained from the family member closest in kinship.

Instrumentation and Monitoring

Each patient had undergone nasotracheal or orotracheal intubation with a standard endotracheal tube (ETT) (Mallinckrodt Laboratories Ltd., Athlone, Ireland) (range 7.5-9.0 mm i.d.). During the study period all patients received a continuous intravenous infusion of fentanyl and propofol at approximately 1 µg/kg/h and 2 mg/kg/h, respectively. Adequate muscle relaxation was achieved by administration of pancuronium bromide in boluses. Patients were connected to a Servo ventilator 900 C (Siemens Elema, Solna, Sweden) and mechanically ventilated in the volume control mode (VCV) with a VT of 8.2 ± 1.9 ml/kg and a respiratory rate (RR) of 15 ± 5 bpm. The inspiratory to expiratory (I:E) ratio was set at 1:1 and the inspired oxygen fraction (FI_O2) was set at 1 throughout the study.

Under bronchoscopic guidance the ETT was advanced to a level 1 to 2 cm above the carina. A 7-fr-i.d. RTC, providing a backward flow through an annular orifice, was slid inside the ETT through an adaptor. The RTC was meant to lie within the distal end of the ETT to avoid any damage to the tracheal mucosa (17). The correct position of the catheter was previously determined by inserting the RTC inside an ETT of the same size in use for the patient.

Pressure at the airway opening (Pao) was measured at the proximal end of the ETT. Peak inspiratory pressure (PIP_tr) and positive end-expiratory pressure (PEEP_tr) were measured in the main airways via a small catheter (1.5 mm i.d.) provided with side holes, inserted through the ETT.

An esophageal balloon catheter (Smartcath, Bicore Monitoring Systems, Irvine, CA) was positioned in the distal third of the esophagus and attached to a pressure transducer to record esophageal pressure (Pes). Pao, tracheal pressure, and Pes were measured with calibrated pressure transducers (Bentley Trantec Inc., Armstrong, CA) and simultaneously recorded on an eight-channel thermal recorder (TA 5000 recorder; Gould Instruments, Cleveland, OH). Flow was measured by a heated pneumotachograph (Fleisch #2) mounted at the Y-piece, connected to a differential pressure transducer (Validyne MP 45; Validyne Co., Northridge, CA) and calibrated with oxygen. Volume was computed by digital integration of the flow signal (see below). The flow, Pes, Pao, and tracheal pressures signals were processed via an analog-to-digital converter (80 samples/s per channel) by an IBM portable PC for storage and later analysis.

In the last six patients studied, bands for respiratory inductive plethysmography (RIP) (Respitrace Plus; NIMS, Inc., Miami Beach, FL) were placed around the chest and the abdomen to detect changes of EELV while switching from one mode of ventilation to another, as well as to help ensure a constant VT (24). RIP was used in the direct corrent-coupled mode; signals were recorded on paper and processed by a PC for later analysis.

A mixing chamber (5 L) attached to the expiratory port of the ventilator allowed continuous sampling of mixed expired CO₂ concentration (Datex Normocap, Instrumentarium Corp., Helsinki, Finland). At the beginning of each study the CO₂ analyzer was calibrated with a gas mixture of known CO₂ concentration (5.3%). The efficiency of the mixing chamber was individually tested in each patient by comparing mixed expired CO₂ concentration with the data obtained from a standard expiratory gas collection procedure.

Arterial and mixed venous blood samples were analyzed at 37° C (Radiometer ABL, Copenhagen, Denmark) and corrected for body temperature. Right-to-left shunt (S/T) was computed from calculated arterial and mixed venous oxygen contents.

ITPV System and Calculations

During ITPV, a continuous flow of humidified oxygen was supplied to the RTC through a flowmeter. Gas flow was delivered through two in-series Conchatherm III temperature-controlled humidifiers (Hudson RCI, Temecula, CA). Pressure proximal to humidifiers and temperature of the two heating columns were continuously monitored. The ventilator was set in the pressure control mode at 0 cm H₂O above the PEEP level (i.e., no flow was delivered). The inspiratory valve remained closed at all times, while the expiratory valve of the ventilator acted as a shutter. When it was closed all flow delivered through the RTC entered the lungs; when it opened, the expiratory VT plus the continuous intratracheal flow from the RTC were exhaled (18). The expiratory valve opening and closure times depended on the RR and I:E ratio, which were checked on the PC recordings of the airway pressures and gas flow tracings. Because during ITPV all continuous flow was delivered directly to the catheter and inspiratory flow from the ventilator equaled zero, the actual VT delivered at each respiratory cycle (VT, ITPV) was proportional to the expiratory valve closure time. Approximating for V_insp  V_exp in the respiratory system, and given an I:E ratio of 1:1, the effective inspiratory VT (VT, ITPV) was obtained from the total expiratory flow (_exp) and computed as follows:

CO₂ was calculated as the product of the measured mixed expired CO₂ (FCO₂) and total minute volume. Physiologic dead space fraction (VD/VT) was calculated from the Enghoff modification of the Bohr equation (25) as follows:

During baseline VCV, static respiratory compliance (Cpl,rs) and total inspiratory resistance of the respiratory system (Raw,rs) were assessed by an occlusion maneuver. Appropriate corrections were made for compressible volumes (approximately 0.7 ml/cm H₂O) (26). Static auto-PEEP (PEEP_i) was measured during the end-expiratory occlusion. Cpl,rs was obtained by dividing the expired volume by the difference between end-inspiratory elastic recoil pressure and end-expiratory occlusion airway pressure (PEEP + PEEP_i). Raw,rs was computed from the pressure tracing measured at the airway opening during an end-inspiratory occlusion, as previously described (26).

During both VCV and ITPV, the difference between end-inspiratory and end-expiratory esophageal pressure (Pes) was measured as an indirect index of inspiratory VT. Mean airway pressure (MAP) at the airway opening and MAP_tr were calculated by time averaging the signals over five respiratory cycles.

Experimental Protocol

After patient preparation, we allowed 60 min of equilibration before collecting the first baseline data set. Measurements were taken before and after the RTC insertion (without flow) to detect changes on respiratory mechanics and/or gas exchange. Measurements obtained after the insertion of the RTC were considered as the baseline (B₀). Thereafter, ITPV was started after switching to the pressure control mode as described, while the intratracheal flow was progressively increased to reach the same VT as during baseline VCV.

To verify baseline reproducibility and stability, each patient was returned to baseline VCV (B₁). Each experimental period lasted 30 to 40 min, after which time data were collected.

PEEP_tr was meant to be kept constant during ITPV. FI_O2 was set at 1 and was kept constant throughout the protocol, as well as RR and the I:E ratio. When available, RIP signals allowed continuous monitoring of inspired VT and EELV. Analysis of recorded data was performed when full stability of VT and PEEP_tr was achieved. At each experimental stage, mixed venous and arterial blood samples were obtained and hemodynamic parameters were collected.

Statistical Analysis

Two-way analysis of variance (ANOVA) for repeated measurements was used to evaluate differences for each variable among B₀, ITPV, and B₁ data sets. Relationships between selected parameters were evaluated by simple regression analysis. Regression equations, when reported, are expressed with the same unit of measurement used in tables or in the text. Data are expressed as means ± SD. Probability values lower than 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Raw,rs before and after the insertion of the RTC (without flow) increased from 14.3 ± 8.6 to 17.4 ± 6.5 cm H₂O/L/s, respectively (p < 0.05). PEEP, Cpl,rs, and gas exchange were not affected by the presence of the RTC (data not reported). Baseline PEEP_i during VCV was 0.43 ± 0.61 (range 0.2-2.0 cm H₂O). The intratracheal flow rate averaged 20.1 ± 3.8 L/ min (range from 14.5 to 25.8 L/min). The working pressure, developed in the gas delivery system, ranged from 350 to 820 mm Hg (540 ± 190 mm Hg). Temperature of the delivered gas was steadily maintained at about 37° C. No adverse events were detected during any experimental stage.

Gas Exchange

Stability in VT between VCV and ITPV stages was achieved (8.2 ± 1.9 and 8.9 ± 1.8 ml/kg, respectively, p = 0.37), and I:E ratio was constant at 1:1. No changes in Pes were observed between experimental periods (Table 2). While VT was constant, VD/VT significantly decreased during ITPV (Table 2) due to a decrease in VD of 65 ± 45 ml (p < 0.01). Consequently, alveolar ventilation (A) increased by 26 ± 12% (from 3.52 ± 0.82 L/min at VCV to 4.54 ± 0.94 L/min during ITPV, p < 0.005). Arterial PCO₂ (Pa_CO2) consistently decreased with an average percentage reduction relative to baseline of 15 ± 4% (range 10 to 20%); accordingly, arterial pH improved (Table 2). Individual Pa_CO2 changes are shown in Figure 1. As expected, the greatest CO₂ decrease was observed in patients with the highest baseline Pa_CO2 levels (Pa_CO2 = 0.29 × Pa_CO2 9.48; r = 0.95, p < 0.001). After return to B₁, Pa_CO2 was slightly but significantly lower than B₀ (Table 2). CO₂ was significantly higher during ITPV and then returned to its baseline value during B₁ (Table 2). Changes in arterial PO₂ (Pa_O2) did not reach the level of significance, although s/T was significantly different among the three stages (Table 2).

View larger version (21K): [in this window] [in a new window]   Figure 1.   PaCO2 at baseline volume-cycled mechanical ventilation (B0), during ITPV and after return to baseline mechanical ventilation (B1) in each patient. Horizontal lines indicate mean values (n = 11) and boxes indicate SD.

Lung Volume and PEEP

When RIP was available, the measured VT was constant throughout the protocol. During transition from VCV to ITPV, as intratracheal catheter flow was started and then progressively increased to match baseline VT, dynamic PEEP_tr decreased. Accordingly, EELV progressively declined (Table 3). Baseline values were restored by increasing the external PEEP at the ventilator by an average of 3.6 ± 2.0 cm H₂O (Figure 2).

View larger version (36K): [in this window] [in a new window]   Figure 2.   Transition from baseline volume-controlled mechanical ventilation (VCV) to ITPV for a representative patient (#9). From top to bottom, pressure tracing at the airway opening (Pao), at the carina (Pawtr), and respiratory inductive plethysmography signal (Vol). After matching baseline VT, intratracheal gas insufflation resulted in a reduction in PEEPtr with a respective decrease in end-expiratory lung volume (EELV). Both were restored by increasing PEEP at the airway opening (arrow).

Hemodynamics

Patients showed no relevant temperature changes throughout the protocol. No major hemodynamic changes were observed. Nevertheless, cardiac output (CO) slightly decreased during ITPV and, accordingly, there was a small but significant reduction in mixed venous saturation (Table 4). Although of little clinical relevance, mean pulmonary arterial wedge pressure () was slightly but significantly lower during ITPV (Table 4, p < 0.05), whereas central venous pressure (CVP) did not significantly change at any stage.

Mean pulmonary arterial pressure () marginally decreased during ITPV reaching the level of significance (Table 4). Regression analysis did not show any significant correlation between changes in and Pa_CO2, pH, or Pa_O2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

General Comments

The main results of this study could be summarized as follows: in a selected group of patients with ARDS, ITPV was able to deliver the entire VT, bypassing the anatomical dead space and providing a significant reduction of Pa_CO2.

In spite of the use of relatively high gas flow rates, the RTC was able to facilitate expiration, as indicated by the need for an increased external PEEP and by a constant EELV as measured by RIP.

Gas Exchange

Our results are similar to those reported with continuous TGI at moderate catheter flow rates (4-6 L/min) in patients with acute respiratory failure (13, 15, 27). More recently, Belghith and colleagues obtained an average Pa_CO2 reduction of approximately 20% at a catheter flow rate of 4 L/min in a group of severely hypercapnic patients with ARDS (19).

In the current study, we enrolled moderately hypercapnic patients with quite high metabolic rates. ITPV yielded a decrease in VD that accounted approximately for the estimated apparatus dead space.

In the setting of acute lung injury, an increased alveolar dead space fraction is expected to reduce the efficacy of intratracheal insufflation techniques. This effect can be partially counterbalanced by the implementation of permissive hypercapnia with higher alveolar CO₂ concentration and a greater VD/VT. In a recent experimental study on oleic acid-injured dogs, Nahum and colleagues evaluated the CO₂ clearance efficiency of expiratory TGI at 10 L/min (28). In their experimental conditions, the authors have shown how, for a given decrease in VD, Pa_CO2 is decreased by different amounts according to baseline Pa_CO2, VD/ VT, and VT. Such a relationship can explain the observed Pa_CO2 reduction in our clinical setting, in which VT was not as low as those implemented in more severe patients with ARDS.

Nahum and colleagues tested the effect of an inverted-tip insufflation catheter compared with a straight catheter with respect to CO₂ removal efficiency (8). They demonstrated that the efficacy of continuous TGI is primarily related to expiratory washout of proximal dead space. They also concluded that a minor contribution to increased CO₂ clearance may be related to a jet effect created by high flow rates through the straight catheter (up to 15 L/min) with turbulent gas mixing distally to the catheter tip.

The RTC we used, with its upward thrust during expiration, may be less efficient with respect to CO₂ clearance, although a recent experimental study conducted on spontaneously breathing animals did not show any difference between the RTC and a straight catheter (29). Nevertheless, this aspect needs further exploration.

Differences in Pa_CO2 between the two VCV stages, B₀ and B₁, might be explained by the fact that interval times elapsed among measurements were not long enough to achieve a complete steady state.

We could not detect any significant change in Pa_O2, although in some patients there was a slight increase in arterial oxygenation. If MAP and mean lung volume, which are the major determinants of arterial oxygenation, are kept constant, Pa_O2 is not expected to change during TGI, as recently reported in a study on oleic acid-injured dogs (30).

Changes in cardiac output were not significant, although a slight decreasing trend could be recognized during ITPV. Explanation for this nonsignificant difference is not immediately obvious. An increase in intrathoracic (i.e., intrapulmonary) pressure can be excluded as the cause of it, as CVP, , and all show either a significant decrease or a decreasing trend (Table 4).

Pressure and Volume Effects

According to previous experimental observations, the turbulence of the flow emerging from the RTC produces a jet that creates a Venturi effect (17, 22). The pressure decrease depends on the following factors: (1) the ratio of the cross-sectional areas of the ETT and the catheter; (2) the gap at the catheter tip where the catheter flow emerges; (3) gas velocity; and (4) duration of expiration.

During constant flow ventilation, dynamic pressure measurements distal to the catheter tip can be influenced by pressure fluctuations related to turbulence and coexistence of inspiratory and expiratory flows in the region of the carina (31). However, mechanisms of flow and pressure distribution during TGI are less known. During continuous TGI, EELV and dynamic PEEP increase in direct proportion to flow and inspiratory time, as recently reported using an argon washout method (32). Nahum and colleagues in their study conducted with an inverted-tip catheter (8), found that tracheal pressure measurements did not accurately reflect changes in lung volume and, consequently, in mean alveolar pressure. They suggested that the inverted jet entrained alveolar gas, accelerating expiratory flow during the early phase of expiration with an increase in expiratory flow-resistive pressure losses. Such a mechanism still remains unclear and, although there might be some similarities with ITPV with respect to the direction of flow during the expiratory phase, it is still unknown how to apply this mechanism to the peculiar design of the RTC. Another recent study reported a decrease in static PEEP measured at end-expiratory occlusion during expiratory-phase TGI with a modified ETT that produced a reverse flow (33).

In general, however, a major problem in TGI clinical experiments, even when low flow rates have been used, has been that EELV invariably increased. However, to maintain a constant EELV we always had to increase external PEEP to compensate for the observed decrease in PEEP_tr. Therefore, the RTC may be a useful tool to prevent hyperinflation during tracheal insufflation at high flow rates, as it was in our case.

Nevertheless, further investigations of these aspects are required, including setups that may allow static mean alveolar pressure measurements.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

In the present study, ITPV effectively reduced dead space ventilation in patients with ARDS. Our preliminary data suggest a potential benefit from the use of the RTC. Some technical concerns may arise in the long-term clinical application of the technique. Particularly, humidification of the delivered flow at such flow rates may be troublesome. Further technological improvement is required to apply the technique with appropriate safety devices and more studies are needed to test its efficacy at different ventilator settings.