Recognition of the potential for ventilator-associated lung injury has renewed the debate on the importance of the inspiratory flow pattern. The aim of this study was to determine whether a ventilatory pattern with decelerating inspiratory flow, with the major part of the tidal volume delivered early, would increase functional residual capacity at unchanged (or even reduced) inspiratory airway pressures and improve gas exchange at different positive end-expiratory pressure levels.
Surfactant depletion was induced by repeated bronchoalveolar lavage in 13 anesthetized piglets. Decelerating and constant inspiratory flow ventilation was applied at positive end-expiratory pressure levels of 22, 17, 13, 9, and 4 cm H(2)O. Tidal volume, inspiration-to-expiration ratio, and ventilatory frequency were kept constant. Airway pressures, gas exchange, functional residual capacity (using a wash-in/washout method with sulfurhexafluoride), central hemodynamics, and extravascular lung water (using the thermo-dye-indicator dilution technique) were measured.
Decelerating inspiratory flow yielded a lower arterial carbon dioxide tension compared to constant flow, that is, it improved alveolar ventilation. There were no differences between the flow patterns regarding end-inspiratory occlusion airway pressure, end-inspiratory lung volume, static compliance, or arterial oxygen tension. No differences were seen in hemodynamics and oxygen delivery.
The decelerating inspiratory flow pattern increased carbon dioxide elimination, without any reduction of inspiratory airway pressure or apparent improvement in arterial oxygen tension. It remains to be established whether these differences are sufficiently pronounced to justify therapeutic consideration.
Key words: Anesthesia, intravenous: ketamine. Lungs: acute respiratory failure; surfactant deficiency. Monitoring: functional residual capacity; gas exchange; hemodynamics. Ventilation: constant inspiratory flow; decelerating inspiratory flow; positive end-expiratory pressure.
ALTHOUGH mechanical ventilators delivering different flow patterns have long been available, it is still unclear whether particular flow patterns have specific advantages in patients with diseased lungs. [1-4]The constant inspiratory flow pattern guarantees minute ventilation but allows airway pressures to increase (volume-controlled ventilation), whereas the decelerating inspiratory flow pattern guarantees the set peak airway pressures (pressure-controlled ventilation).
A number of studies have suggested that pressure-controlled ventilation increases compliance and improves gas exchange while reducing inspiratory airway pressure. [5,6]In our opinion, these reported improvements--indicating recruitment of collapsed lung units and improved alveolar gas exchange--could have been referable to intrinsic positive end-expiratory pressure (PEEP) and mean airway pressure, and therefore were hardly attributable to the inspiratory flow pattern alone.
The decelerating inspiratory flow pattern rapidly increases and maintains inspiratory airway pressure, and the early gas delivery sustains alveolar pressure longer during the ventilatory cycle than constant inspiratory flow, potentially aiding recruitment and improving distribution of inspired gas.
The aim of this study was to determine whether a ventilatory pattern with decelerating inspiratory flow, with the major part of the tidal volume delivered early, would increase functional residual capacity (FRC) at unchanged, or even reduced, inspiratory airway pressures and improve gas exchange. If such an increase occurred, this could indicate recruitment due to the flow pattern alone.
In the current study, tidal volume, inspiration-to-expiration ratio, and ventilatory frequency were kept constant at different PEEP levels while constant or decelerating flow was applied.
Materials and Methods
Thirteen piglets of Swedish country breed, with a mean body mass of 28 (+/-4) kg, underwent bronchoalveolar lavage. Surfactant was removed by a series of ten instillations of 0.9% saline with a temperature of 37 degrees C, each of 1-1.5 l volume, as previously described. [7,8]Pentobarbital (15 mg/kg) plus 0.5 mg atropine was given intraperitoneally, and 15 min later anesthesia was induced [9,10]with 500 mg ketamine and 0.5 mg atropine given intravenously. In addition, 20 mg morphine was given intravenously before tracheotomy, preparation, and introduction of intravascular catheters. Simultaneous with the morphine administration, a 20 mg *symbol* kg sup -1 *symbol* h sup -1 ketamine infusion was started and continued throughout the study. During the remainder of the study, there were no surgical or other painful stimuli. Pancuronium bromide was given as a continuous infusion at 0.26 mg *symbol* kg sup -1 *symbol* h sup -1. The animal's lungs were ventilated with a Servo 300 (Servo Ventilator, Siemens-Elema AB, Solna, Sweden) through an 8-mm outer diameter Mallinckrodt endotracheal tube (Mallinckrodt, Glens Falls, NY). A thermostatically controlled heating pad was used to maintain normal body temperature.
The investigations were performed at the Experimental Laboratories of the Department of Anesthesiology and Intensive Care at the University Hospital in Uppsala, Sweden. The local ethics committee for animal experimentation reviewed and approved the protocol.
Catheterization and Monitoring
Intravascular catheters were placed surgically for measurement of central venous, pulmonary arterial (via the external jugular vein), and aortic pressures (via the carotid artery). Correct placement of catheter positions were confirmed by pressure tracings. Electrocardiogram, heart rate, and all pressures were displayed on a bedside monitor (Siemens Sirecust) and recorded with reference to the mid-thoracic level. Arterial and mixed venous blood gases were measured (ABL 300/OSM III, Radiometer A/S, Copenhagen, Denmark). Carbon dioxide production was recorded by a metabolic monitor (Datex Deltatrac, Datex Instrumentarium, Helsinki, Finland).
Cardiac output was estimated with a COLD System (Pulsion Medizintechnik, Munchen, Germany); details of its method of operation have been published elsewhere. A 4-F fiber-optic catheter with a thermistor was introduced via the femoral artery, and advanced to the descending aorta. The thermistor in the femoral artery catheter connected to the COLD system detects the temperature signal in the descending aorta from which cardiac output is calculated. Right ventricular end-diastolic volume was measured according to a technique described previously. Extrathermal volume and intrathoracic blood volume were measured using the technique of double indicator dilution described in detail elsewhere. [12,13]This technique was first developed for measurement of extrathermal volume. The double indicator consists of 5 mg indocyanine green mixed in 10 ml dextrose 5% in water at a temperature of 5-7 degrees C, and was injected as a bolus into the superior caval vein. The dilution curves for dye and temperature were recorded simultaneously in the descending aorta with the thermistor-tipped fiber-optic catheter. Extrathermal volume was calculated as the difference between the volume accessible to the intravascular indicator, indocyanine green, and the extrathermal volume measured by thermodilution. Intrathoracic blood volume was calculated as the product of cardiac output and the mean transit time of indocyanine green between the points of injection in the superior caval vein and the detection in the descending aorta at the level of the diaphragm. In a study by Lichtwarck-Aschoff et al., the coefficient of variation was reported to be 3.4 +/-2.7% for cardiac output, 3.2+/-1.9% for intrathoracic blood volume, and 11.6+/-4.8% for right ventricular end-diastolic volume. In our laboratory, with triplicate measurements in nine piglets, both prelavage and postlavage coefficients of variation for cardiac output were 4.0+/-3.5%. For intrathoracic blood volume and extrathermal volume, coefficients of variation were 6.14+/-2.2% and 5.9 +/-3.1%, respectively.
Airway pressures were obtained from the digital displays of the ventilator. Before starting the study, the pressure and flow transducers of the ventilator had been calibrated with independent devices. Every morning a preuse functional check was made according to the procedure described in the operating manual of the ventilator.
The static chest-lung compliance was calculated according to the formula chest-lung compliance = tidal volume x (end-inspiratory pressure- end-expiratory pressure) sup -1, but with appropriate modifications to account for the compressible volume. [14,15].
When the end-inspiratory occlusion pressure, and the total PEEP were measured, the hold functions of the ventilator were used for 5 s before the equilibrium values were noted.
For measurements of FRC, serial dead space, and alveolar mixing efficiency the sulfurhexafluoride tracer gas wash-in/washout method was used (for details of the method, see [16,17]). In a study by Larsson et al., with duplicate measurements, the coefficient of variation for FRC was found to be 3% (range 0.2-6.6%). In our laboratory, the coefficient of variation for three sequential measurements in nine animals for FRC, under a broad range of tidal volumes and different flow conditions, was 1+/-0.08%.
Immediately after lavage, the surfactant-deficient lungs were recruited with I:E 1:1, a ventilatory frequency of 25 min sup -1, and tidal volume of 11 ml/kg. The PEEP level was set to produce a peak inspiratory pressure of 50 cm H2O for 5 min.
Immediately after the opening procedure, the inflection point (Pinfl) was determined. The static pressure-volume loop was generated using a constant inspiratory flow of 0.15 l/s, a tidal volume of 1,200 ml, and a fractional inspired oxygen of 1.0. From this pressure-volume loop, Pinflwas determined by inspection. .
After induction of anesthesia and preparation, the animals were placed in the prone position. Lavage was performed as described previously. [7,8,18]After lavage, a recruitment procedure was performed. Thereafter, Pinflwas determined, followed by a second recruitment procedure. Fractional inspired oxygen tension was then set to 0.3 for the studies of constant and decelerating inspiratory flow.
With external PEEP set to 100% of Pinfl(PEEP 22 cm H2O), six animals were assigned to constant, and seven animals to decelerating inspiratory flow as the first ventilatory pattern. After the first measurement, the ventilator was switched to the other pattern, which also was applied at the highest PEEP level of 22 cm H2O. This flow pattern was then continued, and measurements were made with PEEP levels of 17, 13, 9, and 4 cm H2O. The ventilator was then switched back to the flow pattern that had been used as the first ventilatory mode, and measurements were again made with the same PEEP levels. Each setting was applied for 15 min before measurements.
A Servo Ventilator 300 provides volume-controlled ventilation (constant inspiratory flow), as well as pressure-regulated volume-controlled ventilation (PVRC, decelerating inspiratory flow).
The PVRC mode provides the set tidal volume by regulating the inspiratory pressure to a value based on the pressure/volume conditions for the previous breath. By allowing the drive gas pressure to be adjusted between each breath, the set tidal volume is maintained within limits set by the upper pressure limit. Also, the pressure difference between breaths is not allowed to exceed 3 cm H2O, which contributes to the maintenance of the set tidal volume. This is in contrast to conventional pressure-controlled ventilation, in which the set inspiratory pressure is constantly applied during each inspiratory phase and the tidal volume will differ with changing conditions of resistance/compliance.
Calculations and Statistics
All ventilatory volumes and derived parameters are converted to BTPS (body temperature pressure saturated with water) conditions. Indexed values are either related to body mass or to body surface area. Values are given as mean+/-1 SD. A standard statistics package was used (StatView, Abacus Concepts, Berkeley, CA). Differences between the ventilatory settings were evaluated with a t test for all paired differences within each variable. Statistical significance was accepted at P less or equal to 0.01 (**) and P less or equal to 0.001 (***).
Results are presented in Table 1, and in Figure 1,Figure 2,Figure 3. "*" Denotes a significant difference compared to the immediately preceding setting; "dagger" denotes a significant difference for constant inspiratory flow compared to the corresponding decelerating inspiratory flow setting at the same level of PEEP. Indexed values are either related to body mass (tidal volume, end-inspiratory volume, intrathoracic blood volume, extrathermal volume) or to body surface area (cardiac index, stroke index, oxygen delivery index). Blood gas tensions are given in mmHg (1 mmHg = 0.13 kPa).
Tidal volume was kept constant at 11 ml/kg during ventilation with both flow patterns and at all levels of PEEP. During the first two measurements, external PEEP was set to 22+/-3 cm H2O in all animals, which corresponded to 100% of Pinfl.
The two inspiratory flow patterns did not differ regarding either end-inspiratory occlusion airway pressure or end-inspiratory lung volume. With decelerating inspiratory flow and at all PEEP levels, mean airway pressure was higher than obtained with the constant inspiratory flow pattern, and peak inspiratory pressure was lower.
Gas Exchange, Alveolar Ventilation
PaO2did not differ between the two ventilatory patterns. During decelerating flow, PaCO2was 40 mmHg (5.3 kPa) at the highest PEEP level and with constant flow it was 44 mmHg (5.9 kPa). For all levels of PEEP, PaCO2was lower with the decelerating inspiratory flow, which resulted in increased alveolar ventilation.
Hemodynamics and Oxygen Delivery
Hemodynamics, oxygen delivery, and mixed venous oxygen saturation did not differ between the decelerating and constant inspiratory flow patterns.
The major findings were that the decelerating inspiratory flow was no better than constant flow in alveolar recruitment or oxygenation of the blood. The decelerating inspiratory flow increased carbon dioxide elimination and alveolar ventilation. After comments on the surfactant-depleted lung model, these findings are discussed in the following paragraphs.
In the current study, we used the lavage model originally described by Lachmann and coworkers, which our group has adapted and characterized in more detail in previous studies. [7,8,18]This model has the advantage of simplicity, stability, and reproducibility. It produces a twofold to threefold increase in extrathermal volume, an increase in the pulmonary shunt fraction, a decrease in compliance, all related to removal of surfactant by lavage. The shunt fraction increased from 9+/-2% before to 35+/-6.5% immediately after lavage with zero PEEP ventilation. Compliance decreased from 31+/-8 to 13+/-4 ml/cm H2O after lavage.
Peak inspiratory pressure was lower with the decelerating flow pattern. However, this pressure depends on the combined effects of compliance, airway resistance, the flow rate/pattern, the resistance of the endotracheal tube, and the central airways and, hence, is regarded as a poor predictor of alveolar pressure. It is generally agreed that end-inspiratory occlusion pressure and end-inspiratory volume more reliably reflect alveolar conditions. With a constant tidal volume and a set external PEEP, no differences were found in these parameters, and compliance did not differ between the two modes. The expected recruitment capability of the decelerating inspiratory flow pattern in terms of FRC and decreased end-inspiratory occlusion pressure was not supported.
Several studies have suggested that ventilatory modes that limit the peak transalveolar pressure may prevent iatrogenic deterioration of lung function. In this study, however, the decelerating inspiratory flow allowed more efficient ventilation than constant flow, although it failed to reduce the airway pressures. Despite this, the decelerating inspiratory flow may be advantageous in the clinical setting, because this option reduces the likelihood of an inadvertent and uncontrolled increase in alveolar pressure.
Hemodynamics and Oxygenation
Despite higher mean airway pressure with decelerating inspiratory flow, there was no improvement in PaO2. A possible, but still unproven, explanation could be that mean airway pressure above a certain limit does not recruit alveoli further. As FRC remained unchanged, the current results suggest a lack of inspiratory flow-dependent recruitment in the lung model under study. The unaffected PaO2at low levels of PEEP may be explained by the pronounced hypoxic vasoconstriction mechanisms of the porcine lung.
In contrast, in patients with severe respiratory failure (Qs/Qt of 43 to 45%) Abraham and coworkers observed an improvement in Pa sub O2, without any changes in mean airway pressure, when changing from constant to decelerating inspiratory flow. Thus, in these patients, the increase in PaO2with decelerating inspiratory flow was not related to mean airway pressure. We think that this was more likely due to improved ventilation/perfusion matching.
Carbon Dioxide Elimination
For all levels of PEEP, the decelerating inspiratory flow pattern yielded lower PaCO2than constant flow, in accordance with the results of Cole et al. .
The increased efficiency of carbon dioxide elimination found with decelerating flow might well be linked to the early delivery of the major part of the tidal volume. The improvement in carbon dioxide elimination seen with the decelerating inspiratory flow could be due to several factors, [24,25]but the data obtained in this study do not further clarify the responsible mechanisms.
The decelerating inspiratory flow pattern increased alveolar ventilation without reduction of inspiratory airway pressures, or improvement in arterial oxygen tension. It remains to be established whether these differences are sufficiently pronounced to warrant therapeutic consideration.
The authors thank Dr. Anders J. Hedlund, MD, Burn Center, Department of Plastic Surgery, Uppsala University Hospital, for statistical advice.