Lung mechanical behavior during one-lung ventilation

Lung Mechanical Behavior During One-Lung Ventilation George M. Barnas, PhD, Juraj Sprung, MD, PhD, Duckhwan Choi, MD, and Renee Kahn, BS Objective: Switching from two-lung to one-lung ventilation would be expected to have large effects on lung mechanical properties, and these effects may depend on tidal volume and respiratory frequency. These changes in lung mechanics with one-lung ventilation may be similar to pulmonary edema. Deteriorating lung mechanics during pulmonary edema have been attributed to a loss of ventilated lung units. Therefore, changes in lung mechanics caused by one-lung ventilation were measured and compared with those previously seen during pulmonary edema. Design: Prospective study. Setting: Research laboratory. Interventions: After induction of anesthesia, beagle dogs" tracheas were intubated with an endotracheal tube with a bronchial blocker (Univent System Corp, Tokyo, Japan) to apply one-lung ventilation. The proper position of the bronchial blocker during one-lung ventilation was confirmed with a fiberoptic bronchoscope. Measurements and Main Results: Lung elastance {EL) and resistance (RL) were calculated from measurements of airway pressure, esophageal pressure, and airway flow in five anesthetized, paralyzed dogs during sinusoidal forcing at a constant mean airway pressure of 10 cmH20 in a wide range of breathing frequencies (0.2 to 1.0 Hz in intervals of

0.2) and tidal volumes (50, 100, 200, and to 300 mL). Measurements were made before and after the left mainstem bronchus was occluded with the bronchial blocker. During ventilation of both lungs, E, and RE depended relatively little on frequency, and both EL and RE were independent of tidal volume. During one-lung ventilation, EL doubled and, at most frequencies, RL increased; frequency dependences were not increased, and no dependence on tidal volume was observed. Conclusions: The lack of tidal volume dependence in EL and lack of large-frequency dependence in RE during one-lung ventilation are inconsistent with changes induced by severe pulmonary edema. Although decreases in ventilatable lung volume may contribute to increases in lung elastance, other characteristics of mechanical behavior during one-lung ventilation differ from those of pulmonary edema; therefore, other additional mechanisms must be involved in determining lung mechanical properties during severe pulmonary edema.

NCREASED LUNG STIFFNESS is a hallmark of adult respiratory distress syndrome (ARDS). Increased elastance (or decreased compliance) in ARDS has been attributed to pulmonary edema 1and to increased lung surface tension.2 Some investigators have assumed that most, if not all, of this increased elastance results only from the decrease in lung volume available for inflation (ventilatable volume excluding areas with trapped gas) because fluid fills spaces that are normally filled with gas. 3-7 Indeed, some have concluded that mechanical lung tissue properties do not change during edema. 3,5,6 If no such changes occur, then in ARDS the clinician might in effect be managing a functionally small lung rather than a stiff lung of normal dimensions. These concepts have important clinical implications because large tidal volumes and high positive end-expiratory pressures (PEEP) are often used to manage ARDS. If the ARDS lung is indeed only a functionally small lung with normal tissue properties, then applying PEEP and high inflation pressures will hyperinflate the functional lung units and increase the risk of pulmonary barotrauma. These conclusions have led clinicians to target new ventilatory objectives in patients with ARDS: earlier ventilatory priorities to keep normocapnia at the expense of both large tidal volumes and peak inspiratory pressures are changing to permissive

hypercapnia with pressure-limited ventilation, to prevent barotrauma caused by alveolar overdistention. 6,8 Clinically, it is controversial whether specific compliance (1/elastance normalized to lung volume) is normal in patients with acute lung injury,9 which would be the case if tissue properties were unaffected. However, if there are additional changes in lung tissue mechanical behavior with edema other than those caused by decreases in ventilatable volume, consideration of optimal ventilatory parameters during ARDS may be more complex. Lung resistance also increases during pulmonary edema in dogs 1°,11 and humans? and part of this increase may occur because the number of ventilatable pathways is decreased; that is, the total effective airway diameter is low because some airways may be blocked, some lung regions may not be expandable, or both. However, the authors know of no studies that have directly compared the effects of pulmonary edema on resistance with those of decreases in the ventilatable lung volume. An animal model of pulmonary edema, induced by injection of oleic acid, causes many of the morphological and physiological changes that occur during ARDS, especially those in lung mechanical behavior. 12 Although there are some differences between the two conditions, oleic acid-induced edema is widely used to study lung mechanics of ARDS. 12Lung resistance (RI.) and elastance (EL) during sinusoidal forcing at constant mean airway pressure in the physiological ranges of frequency and tidal volume (VT) in dogs injected with oleic acid 1°,1I were previously described. In addition to increasing EL and RI. compared with controls, edema substantially increased VT dependence in EL and frequency dependence in RL. In such a condition, ventilatable volume, measured by helium dilution to exclude trapped gas, decreases about 50% in dogs7--the same as would occur if only one lung was ventilated. In the current study, EL and RL were measured in healthy dogs before and after

I

From the Departments of Anesthesiology and Physiology University of Maryland, Baltimore, MD, and Cleveland Clinic Foundation, Cleveland, OH. Supported by National Heart, Lung and Blood Institute Grants, HL-33009 and HL-44128. Address reprint requests to Juraj Sprung, MD, PhD, Department of General Anesthesiology, E-31, The Cleveland Clinic Foundation, 9500 Euclid Ave, Cleveland, OH 44195. Copyright © 1997 by W.B. Saunders Company 1053-0770/97/1105-001353.00/0

604

Copyright© 1997by W.B. Saunders Company KEY WORDS: airways, elastance, resistance, respiratory mechanics

Journal of Cardiothoracic and Vascular Anesthesia,

Vol 11, No 5 (August), 1997:pp 604-607

ONE-LUNG VENTILATION AND LUNG MECHANICS

occluding the left mainstem bronchus with a bronchial blocker. The data were compared with the authors' previous results in edematous dog lungs 1°,11 to evaluate the degree to which decreases in lung volume may contribute to edema-induced changes in lung mechanics. If it were true that mechanical properties of edematous lungs are caused by reduction of ventilatable volume, then overall mechanical properties of the lungs, including frequency dependency and tidal volume dependency, should be similar to that during one-lung ventilation. This study was designed to test this hypothesis.

METHODS

The study was approved by the Institutional Animal Care and Use Committee. Five female beagles (10.1 _+ 0.4 kg SE) were used. The dogs were anesthetized with 20 mg/kg of pentobarbital given intravenously, followed by a continuous infusion of thiamylal (3 mg/kg/hr) with pancuronium (0.08 mg/kg/hr) to maintain complete muscle paralysis. The trachea was intubated with a 7.5-mm ID endotracheal tube with a movable inflatable bronchial blocker attached alongside the main lumen (Phycon Univent, Fuji Systems Corp, Tokyo). With fiberoptic bronchoscopy, the endotracheal tube tip was positioned 3 cm from the carina; for control measurements, the bronchial blocker remained positioned along the main lumen. Conventional mechanical ventilation was maintained with 100% O2 at 12 breaths/min (Engstrom ER 312, Stockholm, Sweden). The VT was adjusted to keep end-tidal CO2 (as measured with Datex, Capnomac Ultima) between 26 and 36 mmHg. The blood pressure was monitored through a cannula in the femoral artery, and heart rate was determined from an electrocardiogram (Electronics for Medicine monitors M2103B and Mll01C, respectively, White Plains, NY). Arterial O2 saturation was continuously monitored with a pulse oximeter (Nellcor, N-100, Hayward, CA) attached to the tongue. Body temperature was monitored in the pulmonary artery with a thermistor-tipped, triple-lumen catheter and was maintained between 36°C and 38°C with a circulating-water pad. Airway flow was measured with a pneumotachograph (Fleisch No 1) and a differential pressure transducer (Celesco LCVR, Canoga Park, CA) attached to the end of the endotracheal tube. Flow was calibrated with a known rate of 02, and checked with a dry gas meter. Airway pressure (Paw) was measured with a second transducer through a polyethylene tube (PE 200) with several sideholes, inserted through the endotracheal tube lumen until it extended 2 cm beyond the tube tip. Esophageal pressure was measured with a third pressure transducer using a polyethylene catheter attached to a 10-cm thin-walled latex balloon placed in the caudal esophagus. The accuracy of the esophageal balloon placement was verified in both the normal and one-lung conditions with a method modified from Baydur et al, t3 by comparing changes in Paw and esophageal pressure during abdominal compression with the airway occluded. A fourth Celesco LCVR transducer measured transpulmonary pressure (the difference between airway and esophageal pressures). When impedance measurements were to be made, the dog was disconnected from the mechanical ventilator, and a series of sinusoidal volume oscillations were delivered by a piston pump driven by a linear motor. A low oxygen flow to the pump-dog system, which leaked out of the piston housing without affecting flow at the pnenmotachograph, was adjusted to keep mean Paw at 10 cmH20 at each VT used. In preliminary studies, it was determined that this mean Paw was sufficient to allow Pa,~ at the end of expiration to remain positive during forcing with the highest VT used (300 mL) during 0ne-lung ventilation. Thus, with this method, end-expiratory Paw increases with decreasing VT, but mean lung volume is independent of VT. Two-lung (control) ventilation was studied first. After each series of volume forcing, the dog was given three large inflations with a resuscitator bag and reconnected back to the

605

mechanical ventilator. In three of the dogs, VT was first kept constant at 300 mL as the frequency was changed from 0.2 Hz to 0.4, 0.6, 0.8, and 1.0 Hz (ie, 12, 24, 36, 48, and 60 breaths/min, five Cyc!es at each frequency). Measurements at the same series of frequencies were then made at tidal volumes of 100 mL, 50 mL, and 200 mL. The entire set of measurements at each "~T w a s repeated three more times. In the two other dogs, frequency was kept constant at 0.2 Hz, and VT was changed from 300 mL to 100 mL, 50 mL, and 200 m L (at least five cycles at each VT). Measurements at the same series of Vx were repeated at 0.4 Hz and 0.6 Hz. This entire set of measurements was then repeated one more time. Using fiberoptic bronchoscopy, the bronchial blocker was positioned in the left mainstem bronchus. The blocker was inflated until no further increase in peak Paw was observed. Measurements were then repeated in each dog as discussed. Arterial oxygen saturation remained at 100% throughout all the experiments. The first two cycles at each frequency were discarded to avoid transients that may occur on switching. The remaining three cycles were digitized, computer averaged, and analyzed by discrete Fourier transform to give RL and EL, as previously described. 1° Multiple linear regression 14 was used to test for the effect of frequency, VT, and one-lung ventilation on EL and RL. Differences were accepted as significant atp < 0.05. RESULTS

During two-lung and one-lung ventilation, RL decreased with increasing frequency (p < 0.05), but showed no consistent dependence on VT (Fig 1A). The RL during one-lung ventilation was significantly higher than control values at 0.2 through 0.6 Hz (p < 0.05), but not if all values from 0.2 to 1.0 Hz were considered (p > 0.05). During ventilation of both lungs, EL increased slightly as frequency increased (p < 0.05), but showed no consistent dependence on VT (Fig 1B). During one-lung ventilation, neither frequency nor tidal volume consistently affected EL (p > 0.05), but EL was roughly twice as high as control 'values (p < 0.05) at any frequency and VT. DISCUSSION

In the respiratory system, the two lungs are arranged in parallel with respect to airway flow, and EL equals the inverse of the sum of the inverses of the elastances of each lung. If the elastances of each lung are approximately equal, then blocking flow to one lung should double EL. Elastance did approximately double in the current study at each frequency studied (Fig 1). At most frequencies, RL also increased, but not necessarily by a factor of 2 (Fig 1). The change in RL caused by blocking one lung is more difficult to predict because RL has two components, airways resistance and lung tissue resistance. Both are affected in different ways by frequency, VT and ventilatable lung volume, l°m and these effects may, in turn, be differentlY affected by blocking one lung. Predicting the effects on airways resistance is especially problematic because t h e relationship between pressure caused by airways resistance and flow is nonlinear; that is, airways resistance increases with increasing flow. 15 Blocking one lung will double the flow into the remaining lung at any given combination of frequency and VT. By this factor alone, airways resistance does not exactly double. Moreover, airways resistance contains components attributable to laminar, turbulent, and other types of flow. 16 Laminar flow is inversely proPortional to radius raised to the fourth power,

606

O3

BARNAS ET AL

TLV

A

20

0 ..... 0 50ml 0--0 ZX. . . . . ~ 1 0 0 m l A - - A I-I. . . . . r'1200rnll--I 0 . . . . . 0 3 0 0 m l II, ~ , I I ,

i

__J

(5

OLV

15

O4 T

E

10

L)

(D C) cO O3 o3 E)

rY

5

t

I

f

I

I

0.2

0.4

0.6

0.8

1.0

0

0.0

Frequency (Hz) 120

B

I

._1

,:5

4

100

O4 T

80

E

60

l

(2

40

(..) c" E~

20

4--"

CO

0 I,I

0

).0

I

I

0.2

0.4-

0.6

I

I

0.8

1.0

Frequency (Hz)

Fig !" Average resistance (A) and elastance (B) of the lungs of five healthy anesthetized-paralyzed dogs during sinusoidal volume forcing at four different tidal volumes at a Constant mean airway pressure (mean Paw = 10 cmH20), during two-lung ventilation (TLV) and during one-lung Ventilation (OLV). SE, when larger than symbol, indicated by vertical bars; n = 5 for all points except valueS at 0.8 and 1.0 Hz, where n = 3, Some values are slightly offset on the x-axis for the sake of clarity.

whereas turbulent and other flows are not. Halving the radius available for flow by blocking a lung will affect local resistances differently, depending on the type of flow in that area. In addition, areas where flow is laminar during two-lung ventilation may have turbulent flow during one-lung ventilation. The overall RL at any frequency and VT will be determined by these effects on the airways and those on tissue resistance, which, because it is generally coupled to EL, l°,u will increase with

one-lung ventilation. More precise interpretation of the effects of blocking one lung on RL is beyond the scope of this article. Several investigators have concluded from measurements of static elastance in edematous lungs that edema does not alter the intrinsic elastance of the lung tissue, but rather that the edematous fluid simply replaces gas space. 3-6 Can pulmonary edema be considered as merely a lessening of lung volume? There were slight methodological differences between this study and the authors' previous studies of pulmonary edema, 10,1i but it is unlikely that they would affect comparison between the two studies. Although end-tidal CO2 was higher in the study of edematous lungs (39 to 47 mmHg 1°) than in the current study, such differences in CO2 are not likely to affect bronchial muscle contraction. !7 The criterion used for determining the mean Paw level in both studies was identical: mean Paw was kept at the level ttiat resulted in a 0 cmH20 end-expiratory Paw during forcing at 300 mL. This criterion ensured that mean lung volume approximately equaled the functional residual capacity + 150 mL at all amplitudes of forcing. Although mean Paw was slightly higher in the edema experiments (11 to l 4 cmH20), this is an appropriate comparison because mean pressure at any given VT would be elevated during edema in physiologic conditions because of increased EL. These dynamic measurements during one-lung ventilation in the wide range of tidal volumes and breathing frequencies show that the idea that fluid in edematous !ungs replaces gas space is an oversimplification. If ventilatable volume is decreased by roughly half (Fig 1), EL increases to approximately the same levels as during severe oleic acid-induced pulmonary edema in similarly prepared dogs, but EL does not decrease with increas 2 ing VT, as is seen during pulmonary edema. 1°," This dependence On tidal volume was not caused by systematic increases in lung volume (that could cause progressive recruitment of alveoli) as the tidal volume increased because mean lung volume was constant in both studies. The large VT-dependence of EL during edema indicates that, in addition to decreases in ventilatable volume, structural changes may occur in the lung parenchyma that are not detectable in static measurements or that, as suggested by Cook et al, 18 surface film kinetics change. The increased frequency dependence of EL during edema 1° is expected in lungs with flow pathway inhomogeneity,19and such uneveness in the distribution of ventilation has been reported in oleic acid-induced edema. 2°,21 Of course, such heterogeneity was not present during one-lung ventilation, and frequency had no consistent effect on EL in the current study. The Re was much higher during oleic acid-induced pulmonary edema 1°,11 than during one-lung ventilation, especially at the lower frequencies, and displayed a large hyperbolic decrease as frequency increased. Thus, changes in RL caused by the decrease in ventilatable lung pathways again explain only a part of the RL alterations seen during edema. There are several clinical implications of the results of this study. Switching from two-lung to one-lung ventilation will roughly double the airway plateau pressure (determined mostly by EL) needed to ventilate the lung at the same frequency and V:r. Plateau pressure approximates alveolar pressure and can be considered the most relevant index of possible barotrauma and cardiovascular effects. 22 Because in healthy lungs (Fig 1) EL

ONE-LUNG VENTILATION AND LUNG MECHANICS

607

during one-lung ventilation does not change much with frequency, plateau pressure also will not change. Peak airway pressure is additionally determined by RL and flow. That is, the resistive pressure, which adds to the pressure caused by EL, is the product of RL and flow at any time in the respiratory cycle. Because RL decreases with increasing frequency (Fig 1) and flow increases with increasing frequency during one-lung ventilation, peak airway pressures will not tend to increase at high frequency. Therefore, in clinical conditions, in patients with healthy lungs, plateau pressures should be expected to approximately double when switching to one-lung ventilation. If this is not so, then incorrect bronchial catheter placement or pathology in a single lung should be suspected. Moreover, both peak and plateau pressures will remain relatively constant if frequency is increased without changing VT. If this is not the case, then lung pathology is possible. For example, heterogeneity within the lung will cause EL and both peak and plateau pressures to increase as frequency is increased.t9 With ARDS, the considerations for ventilating will be more complex. In addition to increases in EL and RL caused by the edema present, heterogeneity and therefore enhanced frequency dependences in EL and RL should be expected. In healthy lungs, plateau pressures (and thereby, peak pressures) will decrease if VT is decreased during either two-lung or one-lung ventilation. This is because EI. is independent of VT, and the pressure due to

EL will b e proportional to VT. This may not be the case if the lungs are edematous in ARDS. For example, if EL were twice as high at 400 mL VT as at 800 mL VT, changing between the two VT would not change plateau pressure. Thus, regulation of plateau pressures will not be straightforward. This needs to be better studied in ARDS patient s. In conclusion, ventilation of lungs in ARDS should not be considered as ventilation of simply "small lungs." Mechanical consequences of pulmonary edema, likely to be present, need to be considered. Although absolute increases in tissue eIastances during severe pulmonary edema and one-lung ventilation are roughly similar, absence of tidal volume dependence during one-lung ventilation suggests involvement of additional mechanisms during edema. This is further supported by substantially higher lung resistances at low breathing frequencies, and greater dependences on frequencies during edema. Mechanical tissue properties of edematous lung cannot be explained solely by reduction of venti!atable volume, and other mechanisms, such as interstitial swelling, clogging of airways, alveolar flooding, or changes in film kinetics, are likely to be involved. !1

ACKNOWLEDGMENT

The authors thank Dr Colin Mackenzie for his support throughout the experiments.

REFERENCES

1. Malo J, Ali J, Wood LDH: How does positive end-expiratory pressure reduce intrapulmonary shunt in canine Pulmonary edema? J Apply Physio157:1002-1010, 1984 2. Petty TL, Si!vers GW, Stanford RE: Abnormalities in lung elastic properties and surfactant function in adult respiratory distress syndrome. Chest 75:571-574, 1979 3. Eissa NT, Milic-Emili J: Modem concepts in monitoring and management of respiratory failure. Anesth Clin North Am 9:199-218, 1991 4. Gattinoni L, Pesenti A, Avalli L, et al: Pressure-volume curve of total respiratory system in acute respiratory failure: Computed t0mographic scan study. Am Rev Respir Dis 136:730-736, 1987 5. Grossman RF, Jones JG, Murray JF: Effects of oleic acidinduced pulmonary edema on lung mechanics. J Appl Physio148:10451051, 1980 6. Marini JJ, Kelsen SG: Re-targeting ventilatory objectives in adult respiratory distress syndrome. Am Rev Respir Dis 146:2-3, 1992 7. Slutsky AS, Scharf SM, Brown R: The effect of oleic acidinduced pulmonary edema on pulmonary and chest wall mechanics in dogs. Am Rev Respir Dis 121:91-96, 1980 8. Amato MBE Barbas CSV, Medeir0s DM, et al: Beneficial effects of the "open lung approach" with low distending pressure in acute respiratory distress syndrome. Am J Respir Crit Care Med 152:18351846, 1995 9. Mancebo J: PEER ARDS, and alveolar recruitment. Intensive Care Med 18:383-385, 1992 10. Barnas GM, Stamenovic D, Lutchen KR: Lung and chest wall impedances in the dog in normal range of breathing: Effects of pulmonary edema. J Appl Physi01 73:1040-1046, 1992 11. Barnas GM, Sprung J, Kahn R, et al: Lung tissue and airway impedances during pulmonary edema in normal range of breathing. J Appl Physiol 78:1889-1897, 1995

12. Lewis JF, Tobe AH: Surfactant and the adult respiratory distress syndrome. Am Rev Respir Dis 147:218-233, 1993 13. Baydur A, Behrakis PK, Zin WA, et al: A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 126:788-791, 1982 14. Bamas GM, Mills PJ, Mackenzie CF, et ah Dependencies of respiratory system resistance and elastance on amplitude and frequency in the normal range of breathing. Am Rev Respir Dis 143:240-244, 1991 15. Rohrer R: Der Stromungswiderstandin den menschlichen Atemwegen und der Einfluss der unregelmassigen Verweigung des bronchial Systems auf den Atrnungsverlauf in verschiedenen Lungenbezirken. Pflugers Arch Gesamte Physiol Menschen Tiere 162:225-299, 1915 16. Pedley TJ, Drazen JM: Aerodynamic theory, in Macklem PT, Mead J (eds): Handbook of Physiology. Section 3: The Respiratory System, vol III, Part 1. Bethesda, MD, American Physiological Society, 1986, pp 41-54 17. Widdicombe JC: Regulation of tracheobronchiai smooth muscle. Physiol Rev 43:1-37, 1963 18. Cook CD, Mead J, Schreiner GL, et al: Pulmonary mechanics during induced pulmonary edema in anesthetized dogs. J Appl Physioi 14:177q86, 1959 19. Otis AB, McKerrow CB, Bartlett RA, et al: Mechanical factors in distribution of pulmonary ventilation. J Appl Physi01 8:427-443, 1956 20. Modell HI, Cheney FW: Effects of inspiratory flow pattern on gas exchange in normal and abnormal lungs. J Appl Physiol 46:11011107, 1979 21. Schoene RB, Robertson HT, Thoming DR, et al: Pathophysiological patterns of resolution from acute oleic acid lung injury in the dog. J Appl Physiol 56:472-481; 1984 22. Marini JJ, Ravenscroft SA: Mean airway pressure: Physiologic determinants and clinical importance, Part I: Physiologic determinants and measurements. Crit Care Med 20:1461-1472, 1992