Respiration Physiology. 84 (1991) 1-1 i Elsevier RESP 01761

Significance of bulk convection during high-frequency oscillation D.R. Spahn

I,

R. Leuthold 2, E.R. Schmid I and P.F. Niederer 2.

Ilnstitut j~r Aniisthesiologie. Universiti~tsspitai and 21nstitut j~r Biomedizinische Technik und Medizinische lnformatik. Universitiit ZiMch und EidgenOssische Technische Hochschule. ZiMch. Switzerland (Accepted 27 December 1990) Abstract. In 7 anesthetized supine dogs with an anatomic dead space of 115-162 ml, gas transport during high-frequency oscillation (HFO) was investigated at an oscillatory frequency of 15 Hz. Starting with an oscillatory volume effectively delivered to the lungs (VDEL) of 60 ml, measured on line with an ultrasonic airflow meter, VDEL was reduced in steps of 10 ml, down to a VDEL of 30 ml, whereby fresh gas flow rate, airway occlusion pressure and lung volume above functional residual capacity were kept constant. An HFO-circuit without bias tube was used. The volume of endotracheal tube and three port connector, designated as HFO-circuit related rebreathing volume, was 35 ml. Paco2 continuously increased, when VDEL was reduced from 60ml to 40ml and the data fit perfectly to a reciprocal regression (l/Paco, - a + b'VDEL), r 2 ranging from 0.95 to 1.00. Measmed Paco, values at a VDEL of 30 mi (8.26 + 1.77 kPa), however, were significant!v (P < 0.025) higher than Paco2 values predicted by the individual reciprocal regression equations (6.25 + 1.46 kPa). This overproportionate increase in Paco2 due to a reduction of VDEL from 40 ml to 30 ml may be explained by the sudden drop out of bulk convection as a gas transport mechanism between central airways and the surrounding because bulk convection is only possible as long as VDEL exceeds the HFO.circuit related rebreathing volume. Bulk convection therefore is considered an essential gas transport mechanism during HFO and the efficiency o f C O 2 elimination during HFO is critically dependent on the net oscillatory volume, i.e. VDEL minus the HFO-circuit related rebreathing volume and not on the relationship between VDEL and anatomic dead space.

Dead space, in high-frequency ventilation; Gas mixing in lung airways; High-frequency ventilation

Bulk convection and molecular diffusion are the two fundamental gas transport mechanisms during any form of breathing and artificial ventilation (Chang and Farhi, 1973; Scheid and Piiper, 1980; Chang, 1984). The interaction between these two mechanisms generates other modes of gas transport during high-frequency oscillation (HFO) (Chang, 1984) and overall gas transport during HFO is thought to result from direct

Correspondence to: D.R. Spahn, Dept. of Anesthesiology, Duke University Medical Center, P.O. Box 3094, Durham, NC 27710, U.S.A. Reprint requests: D.R. Spahn, Institut fOr Aniisthesiologie, Universitlitsspit~;, Ritmistrasse 100, CH-8091 Ziirich, Switzerland. 0034-5687/91/$03.50 © 1991 Elsevier Science Publishers B.V.

2

D.R. SPAHN et al.

alveolar ventilation by bulk convection, mixing by high-frequency Pendelluft or Out-ofPhase HFO, convective gas transport due to asymmetric in- and expiratory velocity profiles, longitudinal dispersion by the interaction between axial velocity and radial transports due to radial diffusion and/or turbulent eddies and molecular diffusion near the alveolocapillary membrane (Chang, 1984; Schmid, 1984). The true significance of bulk convection to overall gas transport during HFO, however, has never been precisely defined. Recent reports from this laboratory showed, that gas transport can be increased by removing the bias tube (Bush et al., 1989; Leuthold, 1990; Spahn et al., 1990) and that gas transport decreases precipitously in vitro as soon as the oscillatory volume, effectively delivered to the lung surrogate (VDEL) decreased below the volume of the tubes, connecting the lung surrogate with the surrounding (Leuthold, 1990). This volume will be designated for HFO-circuits without bias tube as HFO-circuit related rebreathing volume and consists of endotracheal tube and three port connector. The finding, that overall gas transport precipitously decreases, as soon as VDEL becomes smaller than the HFO-circuit related rebreathing volume, may be taken as evidence, that in vitro bulk convection represents a very important gas transport mechanism during HFO. The aim of the present study was to test the hypothesis, that bulk convection also in rive represents an essential gas transport mechanism during HFO. As the HFOcircuit related rebreathing volume of the HFO-system used in the present study was 35 ml (for details see Methods) gas transport efficiency at a constant oscillatory frequency of 15 Hz was investigated at a VDEL of 60, 50, 40 and 30 ml. In studies on gas transport the oscillatory frequency and VDELneed to be precisely known. The measurement of VDEL, however, is not trivial, if the oscillatory volume is determined e.g. between piston pump and four port connector and not between four port connector and endotracheal tube, VDEL may be overestimated because a part of the oscillatory volume may be lost through the bias tube. Pressure and flow amplitudes vary even along straight tubes (Bush, 1988). The only site, suitable for a precise measurement of VDEL,therefore is the entrance to the endotracheal tube. Besides the site of measurement also the technique of flow measurement is of concern. In most studies on gas transport during HFO where VDEL was not simply approximated by the piston displacement oscillatory volumes were measured with Fleisch type pneumotachographs (Fleisch, 1925). Although it is possible to measure flow precisely with this technique over a wide range of flow and flow characteristics, special flowheads for specific flow characteristics and different ranges of flow are required. The problem, however, using this technology in H FO is, that flow amplitudes and flow characteristics dramatically change within each oscillatory cycle: Starting with zero flow, the gas is accelerated rapidly to a maximum, decelerated to zero flow again and then accelerated and decelerated in the opposite direction, encountering a wide range of flow and flow characteristics. With one single flowhead it therefore seems to be very difficult to precisely measure oscillatory flow with pneumotachographic methods. Furthermore, in pneumotachographic flow measurement the pressure drop across the resistance element usually is only measured at the most lateral capillary and thus assumptions with regard to the flow profile become

G A S T R A N S P O R T IN H I G H - F R E Q U E N C Y O S C I L L A T I O N

3

necessary to determine true volume flow which is of concern, since due to the dramatically changing flow characteristics during each oscillatory cycle flow must be considered largely instationary and thus any assumption with respect to flow profile may be associated with considerable uncertainty. The present study is unique in that for the first time VDELwas directly measured at the entrance to the endotracheal tube with a flow measurement technique independent of the flow characteristics of oscillatory flow. This al|ows to clearly define the ratelimiting parameter in overall gas transport during HFO, a very important goal to better understand gas transport during HFO, explicitly formulated years ago (Chang, 1984), which, however, became possible only recently with the development of a transit time ultrasonic airflow meter with specialized flowhead (Buess et al., 1986; Spahn et al., 1990).

Methods

Seven healthy female purebred beagle dogs (body weight 14-18 kg) were anesthetized and paralyzed with pentobarbital (20-30 mg/kg iv), fentanyl (0.006 mg/kg) and pancuronium (0.12 mg/kg) followed by continuous infusions of propofol (10 mg.kg-~.h-~), fentanyl (1.25 /zg.kg-m.h -j) and pancuronium (0.1 mg'kg-m'h -~) (Spahn etal., 1990). The trachea was intubated with a cuffed endotracheal tube (Hi/Lo, Argyl, 0.9 cm ID with 2 accessory lumina of 2 and 4 mm2 cross sectional area) and the dogs were conventionally ventilated (Servo 900B, Siemens Elema Solna, Sweden) with room air to normocarbia. Body temperature was maintained between 34.5 and 37.0 °C by means of a warming blanket. The dogs were studied while they were lying in a supine position. At the end of the study, residual neuromuscular blockade was reversed with prostigmin (0.05 mg/kg) and atropine (0.02 mg/kg) and the animals were allowed to recover from anesthesia. An HFO-system without bias tube (fig. 1), which has been described in detail previously (Bush et al., 1989; Leuthold, 1990; Spahn etal., 1990) was used in the present study. Briefly, the piston pump (piston diameter 10 cm) was connected to the three port connector with a stiff polyvinylchloride high-pressure tube (ID: 1.2 cm, fgf ~ ':iililiiiiiii!ii~!:!:~:: ii!iii~:.ETT~ fh

pistonpump

:::::::::::::::::::::::::::::::::

!!!!!!!'...... i:;

fi!i!ili~v!::~ ~

,

g'

(--~-a air 8V

Fig. !. Experimental set-up: f g f = fresh gas flow (oxygen), ETT = endotracheal tube (length = 35 cm, ID - 0.9 cm), vs - vacuum source, pcs - pressure control system, sv = safety valves, CT = connecting tube (length - 200 cm, ID = 1.2 cm), air = compressed air. FM = ultrasonic transit time airflow meter. tll = flowhead, allowing on line measurement of the oscillatory volume, effectively delivered to the lungs.

4

D.R. SPAHN

et ai.

length: 200 cm). The second port (ID: 0.6 cm, length: 3 cm) was used as connection to a pressure control system, consisting of a box (volume: 5 L) which was flushed by compressed air at a rate of 60 L/rain (Leuthold, 1990; Spahn et al., 1990). By varying the pressure in this box (range: - 60 to + 60 mbar) mean airway pressure, determined by appropriate filtering from the transient pressure, measured with a 16G catheter (length" 60 cm), protruding 5 cm from the tip of the endotracheal tube (Schmid et aL, 1981), was adjusted to 4 + 0.5 mbar. The third port was attached to the endotracheal tube via a specialized flowhead of an ultrasonic transit time airflow meter (Tuba001, G HG, Medical Electronics, Zurich, Switzerland)(Buess et al., 1986)and the oscillatory volume, effectively delivered to the animal (VDEL) was measured on line. The specially designed flowhead had an internal volume ofonly 8 ml (ID: 1.2 cm, length: 7 cm). Fresh gas flow (fgf) was introduced at the tip of the endotracheal tube (Rossing et al., 1984). Because extremely high Pace2 values were expected at VDEL of 30 mi pure oxygen was used as fgf (100% water vapor saturated) to prevent extreme forms of hypercapnic hypoxeraia which would have endangered the safety of the experimental animals. The fgf rate of 4 L/rain was controlled by a needle valve and was measured by a previously calibrated rotameter. Directly prior and immediately after the experimental period, during which the dogs were ventilated by means of HFO, CO2 production and the anatomic dead space were determined during conventional mechanical ventilation at a frequency of 15 breaths per rain. CO2 production was measured with an automatic respiratory gas analyzer (MMC Horizon, Sensor Medics, Anaheim, CA) and the anatomic dead space was calculated by the Bohr equation (Nunn, 1977): VDannt - VT'

Fetco2- FEco2 Fetco, - Flee 2

where V D a n a t - anatomic dead space [ml], VT -- tidal volume [ml], Fetco2 ffi endtidal fraction of CO2 [ ~o ], FEco, = mixed expired fraction of CO2 [ ~o], and Flee 2 inspired fraction of CO2 [ % ]. In all HFO settings an oscillatory frequency of 15 Hz was used. Four different levels of VDEL were investigated. Starting with a VDEL of 60 ml, the oscillating volume was reduced in steps of 10 ml down to a VDELof 30 ml. At each level of VDEL, 30 tO 90 rain were allowed to reach a steady state in terms of Pace2 and Pao,. Steady state was thereby defined as the situation, in which neither Pace 2 nor Pao, showed a consistent trend in 3 consecutive blood gas analyses, taken in 5 rain intervals (AVL 940, AVL, Schaffhausen, Switzerland). Blood gas analyses, airway occlusion pressure and lung volume above functional residual capacity (FRC) were determined after having reached a steady state at each level of VDELinvestigated. For this purpose, the endotracheal tube was clamped during oscillation and the airway occlusion pressure was recorded. The HFO-system was then disconnected, the endotracheal tube was attached to the spirometer (Wright's Respirometer, Ferraris) and the gas volume of the following passive exhalation was measured (Crawford and Rehder, 1985).

GAS TRANSPORT IN HIGH-FREQUENCY OSCILLATION

5

At each level of VDEL, lung volume above FRC, airway occlusion pressure and body temperature were analyzed with a one way analysis of variance for repeated measure design, using SAS (Version 6.03, SAS Institute Inc., Cary, NC). The measured and the predicted Pace, and Pao, values (for details see Results) were compared with paired t-tests, whereby P < 0.05 was considered to be significant.

Results

The mean anatomic dead space was 143 + 16 ml (mean + SD)(range: 115-162 ml) or 8.83 + 0.62 ml/kg (8.26- 10.1 ml/kg) (BTPS). The anatomic dead space measured just before HFO (145 + 11 ml) was not significantly different from the value, determined immediately following HFO (142 + 21 ml). Mean CO2 production was 86 + 16 ml/min or 5.27 + 0.66 ml. kg- ~. min- ~ (STPD). Mean CO2 production before (87 + 16 ml/min) and =d'ter HFO (85 + 17 ml/min) were identical. There were no statistically significant differences between airway occlusion pressures (4.1 + 0.2 - 4.3 + 0.4 mbar), lung volumes above FRC (147 + 22 - 163 + 21 ml) and body temperature (34.8 + 0.6 - 35. I + 0.6 ° C) at the various VDELinvestigated (60, 50, 40 and 30 ml). Pace, values at a VDEL of 60, 50, 40 and 30 ml are shown in fig. 2. Pace, values at a VDEL of 60, 50 and 40 ml showed a nearly perfect fit to a reciprocal regression (l/Pace, = a + b. VDEL), the individual r2-values ranged from 0.95 to 1.00, i.e. 95-100 % of the observed Pace, course in the range of VDELfrom 60 ml to 40 ml could be explained by the applied reciprocal regression model (Sachs, 1984). In fig. 3, Pace= values and the resulting hyperbolic curve from the reciprocal regression analysis, based upon the dat~ at a VD~L of 60, 50 and 40 ml are depicted for dog nr. 3, the dog with the lowest rS-value. The individual reciprocal regression equations, however, signifi-

12 1013. s......s

0

0 m O.

86-

..,,,, \

420

I

20

30

I

I

40 50 VDEL [ m l ]

60

70

Fig. 2. Pace2 as a function of the oscillatory volume, effectively delivered to the lungs (VDEL). (0"7) individual dogs, ( 0 ) mean values.

6

D.R. SPAHN et al. 12

le

n

m

Q.

8o 4

2(I

3e

4(I

5e

6e

?e

VDEL [ ml] Fig. 3. Paco, ( [ ] ) as a function of the oscillatory volume, effectively delivered to the lungs (VDEL) in dog nr. 3 with hyperbolic curve and 95~o confidence limits, resulting from reciprocal regression analysis (I/Paco, = a + b. VDEL), based upon data of VDEL = 60, 50 and 40 ml (r 2 = 0.95). Due to the minimal scatter of Paco= there are 5 hidden observations.

cantly underestimated Paco: at a VDEL of 30 ml; the predicted Paco~ mean value at a VDEL Of 30 ml was 6.25 _+ 1.46 kPa, the mean of Paco~ values measured at a VDEL of 30 ml was 8.26 _+ 1.77 kPa (fig. 5). Likewise, Pao, readings at a VDEL of 60, 50, 40 and 30 ml are shown in fig. 4. With the exception of dog nr. 6 (r 2 -- 0.80), the Pao: values also exhibited a nearly perfect fit to the same type of reciprocal regression (l/Pao2 = a + b. VDEL), r 2 thereby ranged from 0.94 to 1.00. The individual reciprocal regression equations, however, again failed to predict Pao., at a VDEL of 30 ml in that the predicted value (39.13 + 7.36 kPa) was significantly lower than the measured Pao, of 47.31 + 5.3 kPa (fig. 5).

6050 40 30

m

,o 10

20

I

I

30

40

"

I

I

I

50

60

70

VDEL [ ml] Fig. 4. Pao, as a function of oscillatory volume, effectively delivered to the lungs (VDEL). ([']) individual dogs, ( 0 ) mean values.

GAS TRANSPORT IN HIGH-FREQUENCY OSCILLATION 12

i

7

60

"k'k 10

5O ,0

6 IX.

30

4 2

o04 Q"

20 PRED30 MEAS30 PRED30 MEAS30

Fig. 5. Predicted Paco, and Pao, at a VDEL of 30 ml (PRED30) versus the measured values (MEAS30). ~ ' ~ = significantly different (P < 0.025), ~ = significantly different (P < 0.05) (for details see text).

Discussion

The most important finding of the present study is, that bulk convection is an essential gas transport mechanism during HFO. If VDEL is chosen smaller than the HFO-circuit related rebreathing volume, bulk convection can no longer be effective and overall gas transport efficiency is seriously compromised. COMMENTS ON METHODOLOGY

Measurement of VDF.I~. Direct measurement of VDEL is mandatory during H FO: VDEL may be considerably smaller than the piston displacement because part of the oscillatory volume may be lost through the bias system, VDEL, however, may also be larger than piston displacement due to resonant amplification (Brusasco et al., 1986; Bush, 1988; Spahn et aL, 1988). Because large differences in pressure and flow amplitudes were found at various locations within the HFO-circuit (Bush, 1988) the entrance to the endotracheal tube is considered the only suitable location for the measurement of VDEL as long as a measurement at the tip of the endotracheal tube is not possible. On line measurement of VDEL, however, is not trivial. The ultrasonic transit time airflow meter used in this study has three distinct advantages as compared to Fleisch type pneumotachographs: First, there is no fixed resistance element in the HFO-circuit, which might disturb the true sine wave characteristics of HFO flow; second, in contrast to Fleisch type pneumotachographs, our ultrasonic transit time airflow meter is truly capable of measuring volume flow over the entire flow range encountered in the present study irrespective of flow characteristics with one single flowhead; and third, no assumptions with respect to flow profile need to be made to determine true volume flow as with Fleisch type pneumotachographs where the pressure drop is usually only measured across the most lateral capillary of the resistance element. This is of concern since due to the dramatically changing flow characteristics during each oscillatory cycle

8

D.R. SPAHN et al.

flow must be considered largely instationary and thus any assumption with respect to flow profile may be associated with considerable uncertainty. Studies, where VDEL was approximated by the piston displacement, where 'VDEL' was determined between piston pump and four port connector and not between four port connector and endotracheal tube or where 'VDEL' was determined in the bias tube thus might be very difficult to interpret. In the present study therefore, an attempt was made to accurately measure VDEL as close to the animal as possible with an appropriate flow measurement technique. C02 production. CO 2 production during HFO is considered constant due to the facts, that 1) body temperature during the entire experimental period was found to be constant and 2) a continuous infusion anesthesia with profound muscle paralysis was used and 3) CO2 production was shown to be equal right before and immediately after HFO. In addition, lung volume above FRC and airway occlusion pressure were kept constant at all levels of VDEL investigated. We therefore consider the stepwise reduction of VDEL to be causative for the observed Paco2 and Pao, changes. Anatomic dead space. The anatomic dead space (8.83 + 0.62 ml/kg) measured in the present study in purebred female beagle dogs was larger than reported so far for healthy mongrel dogs of either sex, ranging from 4.8 to 5.5 ml/kg (Severinghaus and Stupfel, 1955; Severinghaus et aL, 1957; Drazen et al., 1979). The lower numbers for anatomic dead space of Drazen et al. (1979), do not include the volume of the tracheotomy cannula used in their experiments. The volume of endotracheal tube and three port connector (35 mi or 2.2 ml/kg), however, is part of the anatomic dead space in the present study. The remaining difference in anatomic dead space (without endotracheal tube or tracheotomy cannula) therefore is reduced to 1.1-1.8 ml/kg, a difference, which might be explained by subspecies differences. The important point in our dogs, weighing 14-18 kg, however, is that in each animal all investigated oscillatory volumes were distinctly lower than the anatomic dead space volume. The overproportionate increase in Paco, between a VDEL of 40 ml and a VDEL of 30 ml (fig. 2, fig. 5) therefore cannot be explained by just having reduced VDEL below the anatomic dead space volume with this final reduction of VDEL by 10 ml but rather indicates, that bulk convection is an essential gas transport mechanism during HFO. In an in vitro study (Leuthold, 1990), performed with an HFO-system similar to that used in the present study, i.e. with an HFO-circuit without bias tube, overall gas transport has been shown to decrease precipitously as soon as VDELwas reduced below the volume of tubes, connecting lung surrogate and surrounding. This finding may be taken as evidence, that in vitro, bulk convection represents a very important gas transport mechanism. The volume of tubes connecting central airways and surrounding, the H FO-circuit related rebreathing volume, consisted in the present study of endotracheal tube and three port connector, its volume was 35 ml. If bulk convection also in vivo is considered an essential gas transport mechanism during HFO, a hyperbolic rise in Paco ., is expected, when VDELis decreased from 60 ml to 40 ml, because gas transport

GAS TRANSPORT IN HIGH-FREQUENCY OSCILLATION

9

provided by this mechanism is directly proportional to the product of oscillatory frequency and net oscillatory volume, i.e. VDELminus the HFO-circuit related rebreathing volume. In each individual dog such a hyperbolic rise in Paco, indeed has been shown in the present study for the range of VDEL from 60 ml to 40 ml (fig. 3). However, as soon as VDEL was decreased below the HFO-circuit related rebreathing volume, Paco, increased significantly more, than would have been expected from the reciprocal regression equation (fig. 5). This may be explained by the sudden drop out of bulk convection, because this gas transport mechanism is only possible as long as VDEL is larger than the HFO-circuit related rebreathing volume. Thus in analogy to conventional mechanical ventilation, where the relationship between tidal volume and anatomic dead space volume is crucial for the efficiency of CO2 removal, the relationship between VDEL and the HFO-circuit related rebreathing volume is critical for CO2 elimination in HFO. This conclusion is strongly supported by the observed Pao2 rise during the stepwise decrease of VDEL. It is important to remember, that the fgf in the present study, introduced at the tip of the endotracheal tube, consisted of pure oxygen and that the pressure regulation box was flushed with room air. At a VDEL of 60 ml, where bulk convection is supposed to be highly efficient, the oxygen, introduced at the tip of the endotracheal tube, directly into the central airway compartment, was efficiently washed out to the same extent as room air from the pressure regulation box was washed-in and a Pao: of 24.76 +_ 3.90 kPa resulted. Decreasing VDEL reduced this gas exchange between central airways and the pressure regulation box and Pao, increased continuously (fig. 4). Except in dog no. 6, who showed only a fair correlation (r 2 = 0.80), this rise again could be described precisely with reciprocal regression equations (r2: 0.94-1.00). The Pao, at a VDEL of 30 ml, however, was considerably underestimated by the individual regression equations (fig. 5). This fact again is considered to strongly support our conclusion, that bulk convection must be considered as very effective during HFO. ¢ Strictly speaking this conclusion is valid only for an oscillatory frequency of 15 Hz, because only this frequency has been investigated in the present study. The agreement with our previous finding from an in vitro study covering a frequency range of 5-30 Hz, various oscillatory volumes and a variety of HFO-circuits (Leuthold, 1990), however, is so close, that there is little doubt, that an extrapolation to a wider range of oscillatory frequencies than only 15 Hz might be valid. 15 Hz was chosen because most of the clinical studies on HFO were done at this frequency (Butler et al., 1980; Schmid, 1984; The HIFl study group, 1989) and because one of the probably most beneficial uses of HFO, the use of HFO for high lung volume artificial ventilation in lung injury models (Kolton et al., 1982; Hamilton et aL, 1983) was also first described at an oscillatory frequency of 15 Hz. The concept, that continuous high volume artificial ventilation improves outcome in a lavage type model oflung injury has been validated very recently, even when exogenous surfactant was included in the treatment (Froese et al., 1990); these studies again were performed at 15 Hz.

10

D.R. SPAHN et al.

Conclusion. The findings of this study confirm that C02 elimination during HFO is possible with VDEL'S considerably smaller than the anatomical dead space volume. Efficacy of C02 elimination, however is seriously compromised if VD~L becomes smaller than the HFO-circuit related rebreathing volume. Hence, bulk convection must be considered an essential gas transport mechanism during HFO.

References Brusasco, V., K.C. Beck, M. Crawford and K. Rehder (1986). Resonant amplification of delivered volume during high-frequency ventilation. J. Appl. Physiol. 60: 885-892. Buess, C. P., P. Pietsch, W. GuggenbQhi and E. Koller (1986). Design and construction of a pulsed ultrasonic air flow meter. IEEE Trans. Biomed. Eng. 33: 768-774. Bush, E.H. (1988). Aerodynamisehe Aspekte der Hochfrequenz-OsziUations beatmung (Diss. ETH Nr. 8727). Master's thesis, Institute for Biomedical Engineering and Medical Informatics, Swiss Federal Institute of Technology and University of Zilrich, Switzerland. Bush, E. H., D.R. Spahn, R. Leuthold, G. Kopacsy, P.F. Niederer and E.R. Schmid (1989). Augmentation of CO2 elimination during high frequency oscillation by removing the bias tube - an in vitro study. J. Biomed. Eng. i 1: 334-337. Butler, W.J., D.J. Bohn, A.C. Bryan and A.B. Froese (1980). Ventilation by high-frequency oscillation in humans. Anesth. Analg. 59: 577-584. Chang, H. K. and L. E. Farhi (1973). On mathematical analysis ofgas transport in the lung. Respir. Physiol. 18: 370-385. Chang, H. K. (1984). Mechanisms of gas transport during ventilation by high-frequency oscillation. J. Appl. Physiol. 56: 553-563. Crawford, M. and K. Rehder (1985). High-frequency small-volume ventilation in anesthetized humans. Anesthesiology 62: 298-304. Drazen, J. M., S. H. Loring, A.C. Jackson, J. R. Snapper and R. H. Ingrain (1979). Effects of volume history on airway changes induced by histamine or vagal stimulation. J. Appl. Physiol. 47: 657-665. Fleisch, A. (1925). Der Pneumotachograph: Ein Apparat zur Geschwindigkeitsmessung der Atemlul~. Pfluegers Arch. 209: 713-722. Froese, A.B., P.R. McCulloch, S. Vaclavik, M. Sugiura, F. Moiler and F. Possmayer (1990). Ventilator pattern influences exogenous surfactant effectiveness in the rabbit. Physiologist 33: AI24. Hamilton, P.P., A. Onayemi, J.A. Smyth, J.E. Gillan, E. Cutz, A.B. Froese and A.C. Bryan (1983). Comparison of conventional and high-frequency ventilation: oxygenation and lung pathology. J. Appl. Physiol. 55: 131-138. The H IF! study group (1989). High-frequency oscillatory ventilation compared with conventi~,nal mechanical ventilation in the treatment of respiratory failure in preterm infants. N. Engi. J. Med. 320: 88-93. Kolton, M., C. B. Cattran, G. Kent, G. Volgyesi, A.B. Froese and A.C. Bryan (1982). Oxygenauon during high-frequency ventilation compared with conventional mechanical ventilation in two models of lung injury. Anesth. Analg. 61: 323-332. Leuthold, R. (1990). Gastransportph~nomene bei der Hochfrequenz-Oszillationsbeatmung (Diss. ETH Nr. 9155). Master's thesis, Institute for Biomedical Engineering and Medical Informatics, Swiss Federal Institute of Technology and University of Ziirich, Switzerland. Nunn, J. F. (1977). Applied Respiratory Physiology. 2nd Edition, Butterworths, London, Boston, Durban, Singapore, Sydney, Toronto, Wellingon, pp. 2 ! 5-220. Rossing, T.H., J. Solway, A.F. Saari, N. Gavriely, A.S. Slutsky, J.L. Lehr and J.M. Drazen (1984). Influence of endotracheal tube on CO:, transport during high frequency-ventilation. Am. Rev, Respir. Dis. 129: 54-57. Sachs, L. (1984). Angewandte Statistik. Springer Verlag Berlin, Heidelberg, New York, Tokyo, pp. 298-306.

GAS TRANSPORT IN HIGH-FREQUENCY OSCILLATION

11

Scheid, P. and J. Piiper (1980). lntrapuimonary gas mixing and stratification. In: Pulmonary Gas Exchange. Edited by J.B. West. Academic Press, New York, vol. I, pp. 87-130. Schmid, E. R., T.J. Knopp and K. Rehder (1981). lntrapulmonary gas transport and perfusion during high-frequency oscillation. J. AppL Physiol. 51: 1507-1514. Schmid, E.R. (1984). High Frequency Pressure Oscillations - A New Method for Mechanical Ventilation of the Lungs (Habilitationsschrift 1982). Ziirich, Juris Druck und Verlag Zilrich. Severinghaus, J.W. and M. Stupfel (1955). Respiratory dead space increase following atropin in man, and atropin, vagal or ganglionic blockade and hypothermia in dogs. J. Appi. Physiol. 8: 81-87. Severinghaus,J.W., M.A. Stupfel and A.F. Bradley (1957). Alveolar dead space and arterial to end-tidal carbon dioxide difference during hypothermia in dogs and in man. J. Appl. Physiol. 10: 349-355. Spahn, D. R., E. H. Bush, E. R. Schmid and P. F. Niederer (1988). Resonant amplification and pressure flow characteristics in high-frequency ventilation. Med. Biol. Eng. & Comp. 26: 355-359. Spahn, D.R., R. Leuthold, E.R. Schmid and P.F. Niederer (1990). Gas transport enhancement in highfrequency oscillation. Respir. Physiol. 82: 29-38.