Cardiovascular Effects of High-Frequency Oscillatory and Jet Ventilation

laboratory illd animal studies Cardiovascular Effects of High-Frequency Oscillatory and Jet Ventilation* Jay H. 'Iraverse, M.D .;t Heikki Koroenranta, M.D.;; E. Merrill Adams, Ph.D.; David A. Goldthwait, B.A.; and Waldemar A. Carlo, M.D.

To determine the effects of mean airway pressure on hemodynamics during high-frequency ventilation, we ventilatecl6ve cats (wt 2.8:1::0.6 kg) using both HFOV (frequencies 3 to 20 Hz) and HFJV (frequencies 4 to 8 Hz) at Paw values ranging from 2 to 12 em U.O. Combinations of frequency and tidal volume that maintained oormocapoia were employed in random order before and after reduction of static compliance of the respiratory system by lung lavage. Heart rate was comparable during both modes of high-frequency ventilation. During both HFOV and HFJV, the cardiac output decreased and PVR increased in normal and surfactant-de&cient cats as Paw was elevated (all p
ventilation may offer several adH igh-frequency vantages over conventional ventilation in patients

with respiratory failure. Previous studies have shown that during high-frequency ventilation, equivalent ventilation may be achieved at lower mean airway pressure (Paw) than that used in conventional ventilation.1 Because elevated Paw adversely affects cardiac output (CO),t-4 the use of high-frequency ventilation at lower Paw may reduce cardiovascular depression. As a result, oxygen transport may be enhanced. High-frequency ventilation may be delivered by a variety of ventilators. The most often used modes of high-frequency ventilation are HFJV and HFOV. Although these ventilators share the property of delivering small tidal volumes at very rapid rates, they ventilate by different mechanisms. Because of this difference in operation, it is plausible that they may produce different degrees of cardiovascular impairment. To see if this occurs, we examined the cardiovascular effects of HFJV and HFOV during controlled changes in Paw, tidal volume, and frequency in cats with normal and injured lungs. *From the Department of Pediatrics and Cardiothoracic Surgery, Rainbow Babies and Childrens Hospital, Case Western Reserve University, Cleveland. tDepartment of Medicine, University of Minnesota, Minneapolis. *Department of Pediatrics, University ofTurlru, Turlru, Finls.nd.

1400

slopes and y-iotercepts. Lung lavage reduced the effect of but did not eliminate it. Changes io ventilatory frequency did not affect cardiac output or PVR. We conclude that the interaction between high-frequency ventilatioa and cardiovascular functioo is largely determined by Paw and compliance and is independent of ventilator frequency and the type of ventilator used. (Chat 1989; 96:1400-04)

Paw

CO= cardiac output; HFJV =high-frequency jet ventilation; HFOV =high-frequency oscillatory ventilation; Paw= mean airway pressure; PEEP=positive end-expiratory pressure; PVR =pUlmonary vascular resistance.

MATERIAL AND METHODS

Five cats (2.8±0.6 kg) were anesthetized with an intraperitoneal injection of chloralose (40 mg/lcg) and urethane (250 mg/lcg). Polyethylene catheters were placed in the femoral artery (PE 90) and vein (PE 190) to measure blood pressure and administer fluids and anesthesia as needed. Following a tracheostomy, an endotracheal tube (3.5 rom ID) was positioned above the carina and then tied in place to prevent air leaks. The animals were then placed on a pressure-limited, time-cycled conventional ventilator (Bourns BP200; Flo1 , 0.25; peak inspiratory pressure, 12 em H10; PEEP, 2 em H10; inspiratory-to-expiratory time ratio, 1:5; Pa~ 4.5 em H10 ; respiratory rate, 12/min). Our surgical approach and preparation of the heart have been previously described.• Briefly, the heart was exposed following a midsternal thoracotomy and retraction of the pericardium. Left atrial pressure was recorded from a saline solution-filled catheter (PE 190) placed in the atrium through the left atrial appendage. Pulmonary artery pressure was recorded from an identical catheter placed through a small purse-string opening in the right ventricular wall. All pressure transducers (Statham P23AC) were referenced to zero at the level of the mid-left atrium. An electromagnetic flow probe was placed around the aorta to continuously measure cardiac output (Carolina Medical Electronics EMI416). The PVR was calculated by taking the pressure difference between the pulmonary artery and left atrium and dividing by the CO normalized to body weight. A multiholed chest tube was placed in the thorax to ree:xpand the lungs and drain the chest following closure. The animals were kept on conventional ventilation for 30 min following surgery and were then randomized to periods ofventilation by HFOV or HFJY. During HFOV (Mera Corp, Hummingbird BM0-20N; bias flo~ 5 Umin; Flo1 , 0.25), the animals were ventilated at combinations of Paw (4, 8, and 12 em H10) and

frequency (3, 6, 12, 16, and 20 Hz) in random order, to prevent carryover efl'ects. At each frequency, the stroke volume generated by the osdllator (20, 10, 5, 3. 75, and 3 ml) was adjusted to produce a constant minute ventilation for each combination. However, the actual tidal volume delivered to the animals at these high frequencies could not be accurately measured. HFJV was administered by a jet ventilator of our own design' that utilized an electronically timed solenoid valve producing intermittent jets of a high-pressure gas mixture (Flo,, 0.25) through a narrow cannula (1.6 mm ID) whose tip was placed approximately 3 mm above the proximal opening of the endotracheal tube. Inspiratory-to-expiratory time ratio was set at 1:3. During HFJv. the animals were ventilated at tidal volumes of 4 and 8 ml, frequencies of 4 and 8 Hz, and Paws of 2, 6, and 10 em H 10. At each ~ an equivalent minute ventilation was produced. Paw was controlled by adjusting the PEEP, which was achieved by placing an adjustable valve in the expiratory limb of the ventilation circuit. During both modes of high-frequency ventilation, values for CO, left abial pressure, and pulmonary artery pressure were recorded 30 seconds after all three parameters reached steady-state. This procedure was repeated until the entire frequency range was covered at each Paw Between Paw settings, the animals were returned to conventional ventilation for brief periods to measure end-tidal Pco. (Beckman LB-2). Arterial blood gas values were measured intermittently. The Pa01 was 113 ± 18 mm Hg before lavage and 112± 10 mm Hg after lavage (Coming 175). The Paco. was 39±4 mm Hg before lavage and 40±2 mm Hg after lavage. The pH ranged from 7.36 to 7.44 throughout the study. To assess the cardiovascular effects of high-frequency ventilation in animals with lung injury characterized by reduced compliance, the above protocol was repeated following saline solution lung lavage. The animal's lungs were lavaged with 30 mllkg of 37"C normal saline solution, • ofwhich 80 percent was routinely recovered. During lung lavage, the peak inspiratory pressure during conventional ventilation was raised to 16 em H.O and the Flo1 was set at 1.0 to ensure normoxemia. Respiratory compliance curves generated by the interrupter technique'' were obtained to document a reduction in lung compliance. This required repeating the lavage two to three times in each cat. Compliance measurements were also repeated at the end of the experiment to ensure stability of the preparation. Following lavage, the Flo. on conventional and highfrequency ventilation was kept at 0.40. Thirty minutes were allowed to achieve cardiovascular stability before the protocol was repeated. All data were recorded on a six-channel recorder (Gould Inc, model 260). Analyses of variance with a Newman-Keuls method and Students t tests were used as appropriate. Compliance curves were &t to the equation \bl=a(1-exp{-b X Palv]) where \bl is lung volume above functional residual capacity and Palv is alveolar pressure measured by the interrupter technique. The exponential ooeflicients (b) were obtained by nonlinear regression (BMDP3R~ They were utilized in conjunction with values for respiratory compliance at maximum lung volume and pressure generated by the interrupter technique to compare statistically the changes in compliance before and after lung lavage. The relationship between CO and Paw during both modes of high-frequency ventilation was linear. Therefore, the cardiovascular effects of the two types of ventilators were statistically compared by slope and intercept. Results are reported as mean±SEM. A value ofp<0.05 was required for statistical significance. The study protocol was approved by the committee for animal research.

REsuLTS Following lung lavage, respiratory compliance decreased in all animals. Static compliance at maximal lung volume decreased from 3.2±0.6 to 2.5±0.6 mV

~ c

·e.....

e

.....

160

e POST LAVAGE - HFOV 0 PRE LAVAGE- HFOV • POST LAVAGE- HFJV ~PRE LAVAGE -HFJV

i_,

140

-,,.I_

:::> Q.

.....

:::> 0

(.)


i5

120

---y

100

a::

s

80

---1

2

4

6

8

10

12

MEAN AIRWAY PRESSURE (cmH20) FIGURE 1. The effect of mean airway pressure (Paw) on cardiac output (CO) during high-frequency jet ventilation (HFJV) and highfrequency oscillatory ventilation (HFOV) before and after lung lavage. CO has been normalized to body weight. The results represent the mean of all frequencies and are expressed as mean:tSEM . CO decreased with increasing Paw with both ventilators before and after lung lavage. No difference in CO was found between ventilators. However, CO was higher in each ventilator in lungs with reduced compliance following lung lavage.

em H20ekg (p<0.05). Similarly, compliance represented by the exponential coefficient of the compliance curves decreased from 0.092 ± 0.026 to 0.061 ± 0.037 em-• (p<0.05). Baseline cardiovascular parameters measured during conventional ventilation were not significantly altered by lung lavage. Baseline CO was 120.8 ± 26.5 mVmin•kg before lavage and 120.2 ± 17.2 mVmin•kg following lavage. Mean arterial pressure was 106 ± 12 mm Hg before lavage and 98 ± 15 mm Hg following lavage. All animals remained hemodynamically stable throughout the duration of the experiment as manifested by blood pressure and CO. During HF]v, CO decreased as Paw was increased from 2 to 10 em H 10 in both normal and lavaged lungs (both p<0.01, Fig 1). This was seen at all frequencies and tidal volumes. Increasing tidal volume from 4 to 8 ml reduced CO at all values of Paw (p<0.05). Reducing lung compliance by lavage mitigated the effect of Paw on CO. At each level of Paw, CO was higher in the lavaged lungs. The CO during HFOV was also adversely affected by increasing Paw. This was seen in both normal and lavaged lungs (both p<0.01, Fig 1). CO was higher at each Paw following reduction in lung compliance by lavage. No effect of frequency on CO was observed during HFOV for frequencies >3 Hz. Varying tidal volumes from 3 to 10 ml had no significant effect on CO. However, large tidal volumes (20 ml) used at 3 Hz significantly lowered CO (p<0.01). With both types of ventilators, the effect of Paw on CO was highly correlated. Although there was a wide variability in CO between animals, this was thought to reftect differences in their baseline hemodynamics. CHEST I 98 I 8 I DECEMBER, 1989

1401

Table 1-SI.ope, Y-lntMcept, mad &greuion CoefficWnt: Cardiac Output.,, Mean AinDay Preuure* Before Lung Lavage

HFOV HFJV p Value

After Lung Lavage

m (mllminokg, em H10)

b (mllmin•kg)

r

m (mllmin•kg, em H10)

b (mllminokg)

r

-3.2::!::0.5 -3.4::!::0.4 0.59

138::!:: 17 137::!::21 0.86

0.94::!::0.03 0.90::!::0.04

-2.9::!::0.4 -3.1 ::!::0.6 0.65

151::!::12 144::!::15 0.14

0.94::!::0.02 0.91::!::0.07

•Data are mean::!:: SEM. HFOV =high-frequency oscillatory ventilation; HFJV =high-frequency jet ventilation.

Every animal in our study exhibited a fall in CO with elevations in Paw. Further, all incremental changes in CO with Paw were statistically significant by the Newman-Keuls method. No difference in slope and intercept was observed between the two modes of high-frequency ventilation before or after lung lavage (Table 1). Both modes of ventilation elicited similar cardiovascular responses to Paw. PVR increased with increasing Paw during both HFJV and HFOV (p
Our results support the findings of previous investigations into the effects of Paw on cardiovascular function during high-frequency ventilation. s.Io In addition, by directly comparing the effects of Paw and frequency during HFJV and HFOV in the same animals, we have shown that similar depression of CO is observed during both modes of ventilation. Hence, our results would indicate that no cardiovascular advantage exists between these two types of highfrequency ventilators.

e.....

!8

·e

0.28

e POST LAVAGE - HFOV 0 PRE LAVAGE - HFOV • POST LAVAGE- HFJV ll PRE LAVAGE - HFJV

0.23

c:

0

0.18

N

%

eu

0.13

~

> 11..

0.08 MEAN AIRWAY PRESSURE (cmH20l

FIGURE 2. The effect of Paw on pulmonary vascular resistance (PVR) during HFJV and HFOV before and after lung lavage. The results represent the mean of all frequencies and are expressed as mean::!::SEM . PVR increased with elevations in Paw during both types of ventilation. Lung lavage reduced the effect of Paw on PVR.

1402

Increasing Paw adversely affected CO with both ventilators regardless of lung compliance. However, increases in Paw appear to be better tolerated when lung compliance is reduced. Increasing Paw increased PVR in both normal lungs and lavaged lungs with reduced static compliance regardless of ventilation mode. However, PVR was lower in lungs with reduced compliance. This may partially explain why the adverse effect of Paw on cardiovascular function is dampened when lung compliance is reduced. Ul We used the method of lung lavage5 to reduce respiratory compliance in our animals to examine the cardiovascular response to elevations of Paw in lung injury. We observed no hemodynamic changes in baseline values for CO, mean arterial pressure, PVR, or heart rate following lavage, even though some of the saline solution was not recovered and was presumed to be reabsorbed by the animal's vascular compartment. Although we used different settings of Paw between ventilators, the range of Paw overlapped over most of the data. Furthermore, the slopes did not change over this range of Paw. Although we could accurately measure tidal volume during HFJY, we could not do so during HFOV. Stroke volumes generated by the oscillator were well controlled, but actual volumes entering the animals airway could not be measured. However, it is likely that increases in oscillator stroke volume will produce parallel increases in tidal volume. Actual tidal volumes are dependent on many parameters, including endotracheal tube size and ventilatory frequency. 12 Stroke volume generated by the oscillator may be dissipated by compression of the piston or lost through the lowpass filter. As a result, stroke volume generated by the oscillator may often overestimate the actual tidal volume.13 The mechanisms by which Paw and PEEP affect cardiovascular function have been well studied in all types of assisted mechanical ventilation. The results of many of these studies indicate that elevations in airway pressure produce changes in cardiac function that are independent of the type of ventilation used, be it high-frequency or conventional ventilation. 8 Elevations in airway pressure increase intrathoracic pressure and lung volume, resulting in diminished venous return and increases in PVR. This reduces

right ventricular stroke volume by effectively decreasing preload and raising afterload, ultimately reducing CO. Changes in contractility are more difficult to assess, but the presence of a negative inotrope has been reported in canine plasma following prolonged use of PEEP. 14 We were unable to assess contractility, but it is unlikely that it was altered by the short periods of Paw elevation employed in this study. PVR invariably increases with elevations in airway pressure secondary to increasing lung volume. 15·16 As alveolar volume is increased by raising Paw, vessels surrounding the alveoli are lengthened and compressed, intrinsically increasing their resistance to flow. 17 The degree of alveolar expansion that occurs as Paw is raised is dependent on lung compliance. If compliance is reduced by lung lavage, then one would expect less vascular compression by the alveoli as Paw is raised. Our results and those of others 11 indicate that PVR is indeed less in lungs with reduced compliance when compared with normal lungs at the same airway pressure. Although no previous studies have directly compared the hemodynamic effects HFJV and HFOV in the same subjects, several studies have separately compared the cardiovascular effects of each ventilator with conventional ventilation in animals and humans. Clinical experience with HFJV and HFOV has been obtained in patients following cardiac surgery 18 and in patients with respiratory 1&-21 and cardiovascular failure. 22 The results of these studies indicate that similar or improved blood gases may be achieved with HFJV at comparable or lower mean and peak airway pressures. However, when Paw is comparable between HFJV and conventional ventilation, little or no improvement in CO is observed. 18·19·22 Several studies have investigated the cardiovascular effects of high-frequency ventilation in animals. Otto et al9 compared the hemodynamic effects of HFJV and conventional ventilation in dogs and found no difference in hemodynamics between the two ventilators in normovolemic animals. However, when the animals were made functionally hypovolemic by adding 15 em H 20 of PEEP, CO was found to be higher in those animals ventilated with HF}v. Chiaranda et al10 also studied the hemodynamic effects of PEEP during HFJV and conventional ventilation in normovolemic dogs. They found that HFJV was associated with significantly better cardiovascular function when compared with conventional ventilation at the same level of PEEP. However, the improvement in CO seen in these two studies may be associated with the use of lower Paw during HF]v, and, therefore, it is likely that any advantage of HFJV is directly related to the extent that Paw is lowered with HFjv.9.18 This observation is supported by Sugihara et al, 23 who found no difference in hemodynamics between HFJV and con-

ventional ventilation at comparable Paw in lavaged piglets. Animal models assessing cardiovascular function during HFOV have also produced similar results.4·24 During HFOV, equivalent tidal volumes are actively removed by the ventilator during expiration. In contrast, the expiratory phase of HFJV is passive. This difference in mechanical action may lead to marked variations in gas trapping between the two ventilators. This was experimentally shown in rabbits, where significantly greater amounts of gas-trapping occurred with HFJV than with HFOV.• In our study, however, these ventilation characteristics did not produce a differential effect on hemodynamics. Our results describe the effects of Paw on CO in cats with normal and surfactant-deficient lungs during two commonly used forms of high-frequency ventilation. They show that the degree of cardiovascular depression incurred during this form of assisted ventilation is dictated by Paw and is independent of the type of high-frequency ventilator used. Although the potential for cardiovascular depression appears to be higher during HFJV than during HFOV due to greater gas trapping, we have found no cardiovascular advantage between these two high-frequency ventilators when Paw is carefully controlled. ACKNOWLEDGMENTS: The authors acknowledge the thoughtful comments by Drs. Richard Martin, Avroy Fanaroff, Michele Walsh, and Kingman Strohl, and the assistance in preparation of the manuscript by Vicki Bancroft and Laureen Staltari.

REFERENCES

1 Carlo WA, Chatbum RL, Martin R1, Lough MD, Shivpuri CR, Anderson Jv. et al. Decrease in airway pressure during highfrequency jet ventilation in infants with respiratory distress syndrome. 1 Pediatr 1984; 104:101..()7 2 Rankin 1S, Olsen CO, Arentzen CE, Tyson GS, Maier G, Smith PK, et al. The effects of airway pressure on cardiac function in intact dogs and man. Circulation 1982; 66:108-20 3 Fusciardi 1. Rouby JJ, Benhamou D, Viars P. Hemodynamic consequences of increasing mean airway pressure during highfrequency jet ventilation. Chest 1984; 86:~ 4 Traverse 1H, Korvenranta H, Adams EM, Goldthwait DA, Carlo WA. Impairment of hemodynamics with increasing mean airway pressure during high-frequency oscillatory ventilation. Pediatr Res 1988; 23:628-31 5 Lachmann B, Robertson B, Vogel 1· In vivo lung lavage as an experimental model of the respiratory distress syndrome. Acta Anaesth Scand 1980; 24:231-36 6 Gottfried SB, Rossi A, Calverley PMA, Zocchi L, MUic-Emili J. Interrupter technique for measurement of respiratory mechanics in anesthetized cats. 1Appl Physiol1984; 156:681-90 7 Carlo WA, Goldthwait DA, Korvenranta H1, Traverse 1H, Chatbum RL. Changes in lung volume and airflow alter intrabreath pulmonary mechanics in normal and surfactant deficiency. Am Rev Respir Dis 1986; 133:A152 8 Thompson WK, Marchak BE, Froese AB, Bryan AC. Highfrequency oscillation compared with standard ventilation in pulmonary injury model. 1 Appl Physiol 1982; 52:543-48 9 Otto CW, Quan SF, Conahan TJ, Calkins JM, Waterson CK, CHEST I 98 I 6 I DECEMBER, 1989

1403

10

11

12 13

14

IS

16

17

Hamerof£ SR. Hemodynamic effects of high-frequency jet ventilation. Anesth Analg 1983; 62:298-304 Chiaranda M, Rubini A, Fiore G, Giron G, Carlon GC. Hemodynamic effects ofcontinuous positive-pressure ventilation and high-frequency jet ventilation with positive end-expiratory pressure In normal dogs. Crit Care Med 1984; 12:7S0-54 Mirro R, Busija D, Green R, LeiBer C. Relationship between mean airway pressure, cardiac output, and organ blood ftow with normal and decreased respiratory compliance. J Pediatr 1987; 111:101-06 Marchalc BE, Thompson WK, Duffty P, Miyald T, Bryan MH, Bryan AC, et al. Treatment ofRDS by high-frequency oscillatory ventilation: a prelbninary report. Pediatrics 1981; 99:287-92 Kolton M, Cattran CB, Kent G, Volgyesi G, Froese AB, Bryan AC. Oxygenation during high-frequency ventilation compared with conventional mecbanical ventilation in two models of lung Injury. Anesth Analg 1982; 61 :323-32 Grindlinger GA, Manny J, Justice R, Dunham B, Shepro D, Hechtman HB. Presence of negative inotropic agents in canine plasma during positive end-expiratory pressure. Circ Res 1979; 45:460-67 Howell JB, Permutt S, Proctor DF, Riley RL. Effect of inftation of the lung on dift'erent parts of pulmonary vascular bed. J Appl Physioll961; 16:71-76 &os A, Thomas LJ, Nagel EL, Prommas DC. Pulmonary vascular resistance as determined by lung inftation and vascular pressures. J Appl Physiol1961; 16:77-84 Fishman AP. The pulmonary circulation. In: Fishman AP, ed.

1404

18

19

20 21

22

23

24

25

The respiratory system. Baltimore: American Physiological Society, 1985:11~20 ~lner JH, Chatburn RL, Carlo WA. ~ntilatory and hemodynamic effects of high-frequency jet ventilation following cardiac surgery. Respir Care 1987; 32:332-38 Schuster MP, Klain M, Snyder Jv. Comparison of high frequency jet ventilation to conventional ventilation during severe acute respiratory failure in humans. Crit Care 1982; 10:625-30 Vincken W. Cosio MG. Clinical applications of high-frequency jet ventilation. Intensive Care Med 1984; 10:275-80 Butler WJ, Bohn DJ, Bryan AC, Froese AB. Ventilation by highfrequency oscillation in humans. Anesth Analg 1980; 59:577-84 Crimi C, Conti C, Bu& M, Antonelli M, de Blasi RA, Romano R, et al. High frequency jet ventilation (HFJV) has no better haemodynamic tolerance than controlled mechanical ventilation (CMV) In cardiogenic shock. Intensive Care Med 1988; 14:35963 Suguihara C, Bancalari E, Goldberg RN, Barrios P, Hehre D. Hemodynamic and ventilatory effects of high-frequency jet and conventional ventilation in piglets with lung lavage. Bioi Neonate 1987; 51:241-48 Mirro R, Tamura M, Kawano T. Systemic cardiac output and distribution during high-frequency oscillation. Crit Care Med 1985; 13:724-27 Bancalari A, Gerhardt T, Bancalari E, Suguihara C, Hehre D, Reifenberg L, et al. Gas trapping with high-frequency ventilation: jet versus oscillatory ventilation. J Pediatr 1987; 110:617-

22