Anesth Analg 2006;103:1213-1218
© 2006 International Anesthesia Research Society
doi: 10.1213/01.ane.0000237400.29668.e6
CRITICAL CARE AND TRAUMA
Section Editor:
Jukka Takala
The Effect of Tracheal Gas Insufflation on Gas Exchange Efficiency
Michael R. Pinsky, MD*
,
Edgar Delgado, RRT
, and
Bernard Hete, PhD
From the *Cardiopulmonary Research Laboratory, Department of Critical Care Medicine, Respiratory Care, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; and Respironics Inc, Murrysville, Pennsylvania.
Address correspondence and reprint requests to Michael R. Pinsky, MD, Department of Critical Care Medicine, University of Pittsburgh, 606 Scaife Hall, 3550 Terrace Street, Pittsburgh, PA 15261. Address e-mail to pinskymr{at}upmc.edu.
 |
Abstract
|
|---|
Transtracheal gas insufflation (TGI) improves gas exchange efficiency, but is associated with hyperinflation, and usually requires ventilator adjustment to compensate for the increased gas flow. Although bidirectional TGI (Bi-TGI) minimizes hyperinflation, it does not preclude the need to reduce tidal volumes to prevent hyperinflation. A flow-compensation system was developed by Respironics (Murrysville, PA) to match TGI flows; however, neither that nor the efficacy of Bi-TGI have been tested in vivo. We tested the hypotheses that flow compensation allows for a constant minute ventilation; Bi-TGI produces less hyperinflation than does unidirectional TGI (Uni-TGI), and endotracheal tube size influences the degree of hyperinflation during TGI. Seven anesthetized intact dogs were studied during positive-pressure ventilation using the Respironics flow compensation system. Measurements were made during steady-state conditions at constant and measured levels of CO2 production. Gas exchange efficiency (assessed by expired gas analysis for dead space) and hyperinflation (measured as an increase in pleural pressure) were compared during Bi- and Uni-TGI and for endotracheal tube sizes varying from 7 to 10F. Bi- and Uni-TGI could be delivered at constant minute ventilation without adjusting ventilatory setting when the flow compensation circuit was present. Uni-TGI produced more hyperinflation than did Bi-TGI with all sizes of endotracheal tube, and hyperinflation was universally present as tube size decreased to 7.5F. We conclude that this new flow compensation system allows for the delivery of TGI without the need for adjustments to the ventilator settings, and that Bi-TGI produces less hyperinflation than does Uni-TGI, even with small diameter endotracheal tubes.
|
Introduction
|
|---|
Tracheal gas insufflation (TGI) is an adjunct ventilatory technique that involves insufflation of gas at varying flows (2–10 L/min) through a catheter positioned above the carina. TGI increases exercise tolerance and reduces dyspnea in spontaneously breathing patients with chronic respiratory disease (1–4). Several physical processes are believed to contribute to this effect, including: 1) reducing anatomic dead space (5,6); 2) decreasing the oxygen cost of breathing (7); 3) washing carbon dioxide (CO2) from the airways during expiration, which increases CO2 elimination efficiency (7,8); and 4) reducing the required inspired tidal volume (VT) and/or minute ventilation (VE) needed in ventilator-dependent chronic obstructive lung disease patients (9) by decreasing the dead space to tidal volume ratio (VD/VT) in a flow-dependent manner (9,10). These benefits reduce the number of episodes of acute respiratory failure requiring intubation and assisted ventilation in patients with obstructive and restrictive lung disease (5,6).
However, TGI carries the risk of hyperinflation because it increases airway pressure and, in its simplest form, delivers gas throughout the ventilatory cycle, thus impeding exhalation, promoting dynamic hyperinflation or auto-positive end-expiratory pressure (PEEP), and adding additional gas to the VT during inspiration (11). TGI-induced hyperinflation may minimize or reverse the beneficial effects of TGI on gas exchange efficiency. Several modifications of TGI have been used to address flow-dependent hyperinflation. Continuous TGI can be delivered either antegrade (Uni-TGI) or bidirectional (Bi-TGI) (12). Uni-TGI causes hyperinflation in a flow-dependent fashion, requiring manual reductions in extrinsic PEEP to offset the obligatory hyperinflation (13,14). This external PEEP reduction approach is difficult to accomplish at the bedside because the expiratory airway occlusion estimation of intrinsic PEEP cannot be made during continuous flow conditions. Importantly, we (12) have shown that Bi-TGI, by diverting 40% of the flow retrograde and 60% of the flow antegrade, results in no increase in airway pressure, even at high TGI flows. However, the impact of Bi-TGI on gas exchange efficiency has not been studied. Furthermore, it is unknown to what extent endotracheal tube diameter limits the ability of Bi-TGI to induce intrinsic PEEP. However, with all continuous TGI systems, ventilator-derived VTs need to be reduced to offset the inspiratory phase TGI flow (13).
A novel TGI system developed by Respironics (Murrysville, PA) and validated in our bench laboratory (12) washes out VD but uses a nozzle that delivers gas at the distal port in a bidirectional fashion, such that airway pressure does not increase despite TGI at a flow of up to 10 L/min. Furthermore, the flow circuit has an intrinsic leak compensation that bleeds circuit gas at a rate equal to the TGI flow, thus allowing delivered VT to equal ventilator-defined tidal volume. It is not known, however, if this system improves gas exchange efficiency and delivers an adequate VT when used in vivo and interfaced with a standard commercially available ventilator.
Thus, we examined the effects of Uni-TGI and Bi-TGI on ventilatory efficiency as defined by a decrease in PCO2 for a constant VT, VE and CO2 production (VCO2) (i.e., gas exchange efficiency), their differential impact on hyperinflation as endotracheal tube size varied, and finally, the accuracy of the leak compensation circuit to maintain VE constant with or without TGI in an acute anesthetized canine model.
|
METHODS
|
|---|
The studies described in this report were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh and were in compliance with the Animal Welfare Act and the National Research Council's Guide for the Care and Use of Laboratory Animals. Seven mongrel dogs (weight 20.5 ± 0.8 kg) were studied after a 12-h fast, during which time they were allowed water ad libitum. General anesthesia was induced by a bolus infusion of pentobarbital (30 mg/kg IV) and the animal's trachea was intubated with a 10F Portex endotracheal tube. The animals were placed supine on mechanical ventilatory support using a Puritan-Bennet 7200 ventilator (Nellcor Puritan Bennet, Pleasanton, CA) with VT set at 8–10 mL/kg, or a minimum of 250 mL, room air Fio2 and a ventilatory frequency to produce a baseline Pco2 of between 35 and 40 mm Hg. A triple lumen IV catheter was inserted via a femoral site for infusion of drugs and IV fluids. Anesthesia was provided by a continuous infusion of pentobarbital (0.1 mg ·kg–1·h–1). Supplemental boluses of pentobarbital (50 mg) were infused as necessary to maintain a light plane of anesthesia, as defined by the absence of an eyelid reflex, spontaneous movement, hypertension, and tachycardia. The dog was placed on a heating pad that was used continuously to maintain a core temperature between 37 and 38°C. Mixed expired CO2 and VE were measured from timed 2-min expired gas collections (20 L Douglas bag and super-syringe).
A fluid-filled femoral arterial catheter was inserted via the femoral site into the descending aorta to monitor arterial blood pressure and to withdraw blood samples for blood gas analysis (Radiometer ABL30, Copenhagen, Denmark). A flow-directed balloon-tipped pulmonary artery catheter equipped with a fast-response thermistor for continuous measurement of pulmonary arterial pressure, cardiac output, and mixed venous O2 saturation (SvO2) (Vigilant, Edwards LifeSciences, Irvine, CA) was inserted via an external jugular vein. From a sub-xiphoid approach, a small incision was made in the left thorax at the 6–7 intercostal space mid-axillary line and a (3 x 0.5)-cm thin-walled latex air-filled balloon catheter was inserted into the left pleural space under 10 cm H2O PEEP, so that no pneumothorax would develop. The balloon catheter was placed at the mid-thoracic level. The pressure–volume history of the balloon catheter was determined by measuring balloon pressure at end-expiration with sequential 0.2 mL increments of air. The volume at which the end-expiratory pressure started to progressively increase defined the unstressed volume of the catheter. After determining the balloon catheter's unstressed volume, 0.5 mL less air was inserted into the balloon and the subsequent pressure values were used to monitor pleural pressure (Ppl). We used increases in end-expiratory Ppl as a measure of hyperinflation. After the surgical procedures the animals were allowed to recover for 15 min to return to hemodynamic and gas exchange stability, as defined by the lack of tachycardia, hypo- or hypertension, or changes in end-tidal Pco2 of <2 mm Hg.
Protocol
The experimental setup consisted of noting the effects of different forms of positive-pressure ventilation on gas exchange and hemodynamics after a 15-min stabilization period. In the first three animals, a baseline condition was established with standard positive-pressure ventilation using a 8–10 mL/kg VT delivered in a decreasing inspiratory flow pattern and frequency defined by end-tidal Pco2 or arterial blood gas. The TGI system used a bidirectional TGI catheter (diverting 70% of the flow retrograde and 30% of the flow antegrade) and 10 L/min of TGI flow (Bi-TGI) with a 10 L/min flow leak compensation circuit superimposed on the baseline ventilatory state. No Ppls were obtained on these first three dogs. For the remaining four animals, we added an antegrade-only TGI catheter (Uni-TGI) step, separated by a baseline no-TGI interval using a Latin Square design to account for sequencing effects. The Uni-TGI catheter had a 0.1 in. inner diameter at the flow exit, and both the Bi-TGI and the Uni-TGI catheters had the same outer diameter of 0.125 in. Also in these final four animals, the Vt was set no lower than 250 mL to keep within the specifications of the TGI system leak compensation circuit. In practice, that meant that the ventilator-set Vt increased to 11 mL/kg. The study sequence was baseline, Bi-TGI, baseline, Uni-TGI, baseline. In the last three dogs, we then assessed the effect of baseline versus Bi-TGI versus Uni-TGI on airway pressure and Ppl as the size of the endotracheal tube was varied across five tube sizes (10, 9, 8, 7.5, and 7F). Each sized endotracheal tube was switched in a random order among animals and all measurements of Ppl and airway pressure were performed only after measured variables stabilized (usually after 40 s of ventilation).
Measurements
End-expiratory intravascular pressures, Ppl, SvO2, and cardiac output (triplicate iced D5W bolus injections) were measured and arterial and mixed venous blood gas were analyzed at the end of each step once mixed expired CO2 displayed a constant value for >5 min and at least 15 min had elapsed in the study interval. This state defined CO2 production stability, such that measures of PaCO2 for a constant VT (VE)-frequency reflected Vd/VT.
Analysis
All mathematical manipulations were done with MATLAB (version 12, Mathworks, Natick, MA). Comparisons among and between animals with varying TGI (treatment) or endotracheal tube size (condition) were made by repeated measurement analysis of variance, with a post hoc Student–Neuman–Keuls test performed for all comparisons judged to be statistically significant. Significance was set at P < 0.05. Comparisons of all cardiovascular variables were made between baseline, Bi-TGI and Uni-TGI to identify the hemodynamic effects of TGI. Increases in end-expiratory Ppl were inferred to reflect hyperinflation. VE was measured for each run to document system stability in delivering bias gas flow. End-tidal Pco2 or mixed expired blood gases were compared between Bi-TGI and Uni-TGI to identify any differences in gas exchange efficiency and to calculate Vd (Vd/VT). PaCO2 was measured at each step, and if VE was unchanged, decreases in PaCO2 were inferred to reflect improved gas exchange efficiency. Accuracy of the leak compensation circuit was measured by comparing the ventilator-defined volumes with direct measures of volume from expired gas collected in the Douglas bag and measured by super-syringe.
|
RESULTS
|
|---|
The animals remained hemodynamically stable throughout the duration of the study, without changes in mean arterial blood pressure, cardiac output, or SvO2. Table 1 depicts the effect of the TGI system on gas exchange in the seven animals. For the group as a whole, the mean decrease in PaCO2 with Bi-TGI was 13.9% (n = 7), and with Uni-TGI it was 12.1% (n = 4). VE, as measured by the ventilator between TGI conditions and baseline, was within the ±20% specification of the TGI system, except in two animals (animals 1 and 4) (Tables 1–3). The mean VE, as measured by the ventilator for baseline and Bi-TGI, was 2.7 and 2.8 L/min, respectively. For baseline versus Uni-TGI the mean VE was 2.8 and 2.6 L/min, respectively.
Calculated Vd/VT Using Ventilator-Derived and Directly Measured Vt Data
Vd measures using the expired gas analysis are depicted in Table 2. Mean Vd/VT decreased for both Bi-TGI and Uni-TGI when compared with baseline (P < 0.05). When Vd/VT estimates were made during TGI conditions using ventilator-derived estimates of Vt, the resultant data had some variability, but tended to follow the same trend as expired gas analysis (Table 3).
Effects of Uni- and Bi-TGI on End-Expiratory Ppl As an Estimate of Hyperinflation
Figure 1 displays effects of Bi-TGI versus Uni-TGI on end-expiratory Ppl measured in four animals using an 8F endotracheal tube. Uni-TGI displayed higher end-expiratory Ppl values than either apnea or Bi-TGI in 3 of 4 animals, whereas the fourth animal showed no changes in Ppl among any of the conditions. Figure 2 displays the effect of Bi-TGI and Uni-TGI on hyperinflation, as represented by end-expiratory Ppl in different-sized endotracheal tubes. Bi-TGI and Uni-TGI increased end-expiratory Ppl 13.4% and 28.8%, respectively (–2.1 to –1.8 to –1.4 mm Hg), as compared with baseline. However, with all sized endotracheal tubes end-expiratory Ppl was always higher during Uni-TGI than during Bi-TGI.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 2. Effect of endotracheal tube size on the degree of end-expiratory pleural pressure (Ppl) during intermittent positive pressure ventilation (Base 1 and Base 2) and both Uni- and Bi-TGI.
|
|
|
DISCUSSION
|
|---|
Effect on Gas Exchange
As previously reported (9), TGI enhances gas exchange efficiency by promoting removal of CO2 from the anatomic dead space. We found that, with the exception of one animal in Bi-TGI mode (animal 4, Table 1), TGI decreased PaCO2 for a constant VE. In that animal, during the baseline condition, the ventilator read a lower VE than all other conditions, including the return to baseline at the end of all TGI conditions. Potentially, a leak in the system caused the VE to decrease, therefore causing the 25.1% change in Bi-TGI from baseline. If we use the last baseline (return to baseline condition), then the % difference would only be +5.8% from baseline, therefore meeting system specifications.
Importantly, when more subtle uses of ventilator-derived variables are needed, as with the calculation of Vd/VT from end-tidal CO2, Paco2, and Vt, the estimated Vt values are somewhat more variable than those obtained from the Douglas bag. Since actual Vd/VT decreases with TGI, these data suggest that the flow-compensation value used to match the 10 L/min TGI flow may need some adjustment to support ventilation at the lower limits of delivered Vt.
Both modes of TGI were associated with reductions in PaCO2 relative to baseline, decreasing 13.9% and 12.1% with Bi-TGI and Uni-TGI, respectively. Similarly, ventilator-measured VE values were similar between baseline, Bi-TGI, and Uni-TGI (2.7, 2.7, and 2.6 L/min, respectively). Thus, the delivery system appears to effectively compensate for the TGI gas flow.
The secondary purpose of this study was to test the Respironics TGI system to see if it allowed for effective TGI delivery using Bi-TGI without hyperinflation. Three aspects of this system were studied. First, the ability of Bi-TGI to prevent an otherwise common hyperinflation, Second, the ability of the circuit extrinsic leak to compensate for TGI flow in delivering a normal Vt, and third, the impact of endotracheal tube size on the propensity to develop hyperinflation. As shown in Figures 1 and 2, Bi-TGI was associated with less hyperinflation, as measured by increases in end-expiratory Ppl, than Uni-TGI for the same degree of CO2 removal. Furthermore, the leak compensation circuit matched TGI additional flow without increasing either tidal breaths or end-inspiratory airway or Ppl to tolerance levels of ±12%. Although no specific limits on accuracy have been defined, these limits seem acceptable for clinical practice, since Ppl did not vary.
Our data are consistent with our previous bench study in an in vitro model (11), where Bi-TGI produced minimal to no hyperinflation as compared with baseline measurements. Between that study and the present one, the percentage of forward versus retrograde flow from the Bi-TGI catheter was changed from 60–40 to 70–30. Bench studies demonstrated that this flow diversion resulted in a greater airway opening zero pressure effect (balanced Venturi effect). However, the changes in airway pressure were minor, and probably did not affect our ability to compare these data to our previous bench study. Evaluation of the present study demonstrates that Uni-TGI produced hyperinflation even in large endotracheal tubes (sizes 
8F), whereas Bi-TGI did not. In addition, in tube sizes 7 and 7.5F, Bi-TGI does induce a small amount of hyperinflation. A probable explanation for the development of some hyperinflation with small tube size is the limitation on cross-sectional area inside the endotracheal tube caused by the presence of the TGI catheter, since hyperinflation was not seen with Bi-TGI when larger tubes were used.
Limitations
There are several limitations of this study. First, we studied normal dogs without evidence of airway or interstitial disease. The extent to which increased airway resistance or changes in lung compliance may alter the impact of gas exchange efficiency with Bi-TGI is unknown. However, all our previous clinical studies were in patients with either chronic obstructive lung disease or acute respiratory failure (2–4,8,9). In general, the beneficial effects of Uni-TGI in those settings were always greater with a greater degree of abnormal gas exchange at baseline. Thus, our present finding should extrapolate well to the clinical setting. Still, clinical trials are indicated before Bi-TGI can be used as standard respiratory support. Second, we examined the effect of Uni-TGI and Bi-TGI for only 15 min per step. Although our animals rapidly reached steady-state, it is not clear if a prolonged benefit would be realized with this novel approach. Again, however, prolonged use TGI was its original use, and documented sustained benefits in patients over months (2). Finally, the number of animals studied was small (n = 7), thus our ability to test statistically for other effects of Bi-TGI that may have been present but not yet statistically significant due to the small sample size is limited. Still, we did see that Vd/VT significantly decreased with TGI, and we do not think that this beneficial effect would disappear with an increased sample size.
Clinical Implications
Ventilation accomplishes two goals: oxygenation of the mixed venous blood and off-loading of venous CO2 into the expired air. CO2 excretion is a function of both the CO2 load delivered to the lungs and the alveolar ventilation. However, when lung disease is also present, the amount of lung ventilated but not perfused (i.e., dead space or Vd) increases. One technique used to maintain adequate alveolar ventilation as Vd increases is to also increase Vt. However, large Vt ventilation may induce local lung injury through hyperinflation and barotrauma, and can induce a systemic inflammatory response in patients with preexisting acute lung injury.
Minimizing VTs in subjects with acute lung injury decreases mortality. However, an unfortunate common side effect of Vt-limited ventilation in patients with acute lung injury is alveolar hypoventilation resulting in progressively increasing PaCO2 levels, often referred to as "permissive hypercarbia." Hypercarbia causes marked respiratory distress in the intubated and ventilated patient, requiring increased use of sedation, with its own set of problems and complications. If Vt-limited ventilation could be delivered without hypercarbia, as we have demonstrated using Bi-TGI, then greater patient comfort and less morbidity should be realized in this cohort of high-risk, high-cost patients. The flow compensation aspect of the ventilatory circuit becomes a central part of this management because it maintains Vt at a preset level despite TGI, making TGI safer and easier to use. In addition to being easier to use, Bi-TGI reduces dynamic hyperinflation (auto-PEEP), with its associated increased Vd and cardiovascular effects. Thus, the use of Bi-TGI with flow compensation appears very promising for the adjuvant management of the ventilator-dependent patient.
|
SUMMARY
|
|---|
These data demonstrate that TGI increased gas exchange efficiency in this acute anesthetized canine model (n = 7), when assessed by changes in PaCO2 for a constant VE. However, only Bi-TGI, and not Uni-TGI, consistently prevented hyperinflation, as assessed by increases in end-expiratory Ppl (n = 4). Furthermore, when different-sized endotracheal tubes were used, Uni-TGI consistently induced hyperinflation, with the greatest increases in end-expiratory Ppl observed with the smaller-sized tubes, whereas Bi-TGI produced minimal hyperinflation, and then only with the smallest-sized endotracheal tubes. Finally, estimates of Vd/VT using assumed VTs from the ventilator during TGI displayed some variability, over-estimating or under-estimating delivered Vt at these small tidal volume levels, although these calculated Vd/VT data followed a similar trend to the indirectly measured values (Douglas bag). Using Douglas bag measures, TGI consistently reduced Vd/VT in these animals.
|
Footnotes
|
|---|
Accepted for publication July 29, 2006.
Supported by Respironics Inc, Murrysville, Pennsylvania.
|
REFERENCES
|
|---|
- Heimlich HJ, Carr GC. Transtracheal catheter technique for pulmonary rehabilitation. Ann Otol Rhinol Laryngol 1985;94: 502–4.[ISI][Medline]
- Hoffman LA, Dauber JH, Ferson PF, et al. Patient response to tracheal oxygen delivery. Am Rev Respir Dis 1987;135:153–6.[ISI][Medline]
- Wesmiller SW, Hoffman LA, Sciurba FC, et al. Exercise tolerance during nasal cannula and transtracheal oxygen delivery. Am Rev Respir Dis 1990;141:789–91.[ISI][Medline]
- Christopher KL, Van Hooser DT, Jorgenson SJ, et al. Preliminary observations of transtracheal augmented ventilation for chronic severe respiratory disease. Respir Care 2001;46:15–25.[Medline]
- Bergofsky EH, Hurewitz AN. Airway insufflation: physiologic effects on acute and chronic gas exchange in humans. Am Rev Respir Dis 1989;140:885–90.[ISI][Medline]
- Hurewitz AN, Bergofsky EH, Vomero E. Airway insufflation. Increasing flow rates progressively reduce dead space in respiratory failure. Am Rev Respir Dis 1991;144:1229–33.[ISI][Medline]
- Benditt J, Pollock M, Roa J, Celli B. Transtracheal delivery of gas decreases the oxygen cost of breathing. Am Rev Respir Dis 1993;147:1207–10.[ISI][Medline]
- Nahum A, Burke WC, Ravenscraft SA, et al. Lung mechanics and gas exchange during pressure-control ventilation in dogs: augmentation of CO2 elimination by an intratracheal catheter. Am Rev Respir Dis 1992;146:965–73.[ISI][Medline]
- Hoffman LA, Miro AM, Tasota FJ, et al. Tracheal gas insufflation: limits of efficacy in adults with acute respiratory distress syndrome. Am J Respir Crit Care Med 2000;162:387–92.[Abstract/Free Full Text]
- Scott GC, Hinson JH, Scott RP, et al. The effects of transtracheal gas delivery on central inspiratory neuromuscular drive. Chest 1993;104:1199–202.[Abstract/Free Full Text]
- Delgado E, Hoffman LA, Tasota FJ, Pinsky MR. Monitoring and humidification during tracheal gas insufflation. Respir Care 2001;46:185–92.[Medline]
- Delgado E, Hete B, Hoffman L, et al. Effects of continuous, expiratory, reverse, and bi-directional tracheal gas insufflation in conjunction with a flow relief valve on delivered tidal volume, total positive end-expiratory pressure, and carbon dioxide elimination: a bench study. Respir Care 2001;46:577–85.[Medline]
- Miro AM, Hoffman LA, Tasota FJ, et al. Tracheal gas insufflation improves ventilatory efficiency during methacholine-induced bronchospasm. J Crit Care 1997;12:13–21.[ISI][Medline]
- Miro AM, Lutz J, Pinsky MR, et al. Auto-peep during tracheal gas insufflation: testing a hypothetical model. Crit Care Med 2000;28:3474–9.[ISI][Medline]
Top
Abstract
Introduction
