High-Frequency Jet Ventilation

High-Frequency Jet Ventilation* A Prospective Randomized Evaluation GrazianoC. Carlon, M.D.,EC.C.l?;tWiUiamS.Howland, M.D.,EC.C.l?;S C o b Ray, R.R.T.;§ Saul Mwdownik, M.E.E.:llJoyce l? G d f i n , R. N., M. S.N., C.C.R.N.;Tand Jeffrey S . Groeger, M.D.**

Three hundred nine patients were randomly allocated to two ventilatory protocols; 157 patients were supported with a volume-cycled ventilator (VCV) (Bear Medical BEAR 1) and 152with a high-frequency jet ventilatory (HFJV)developed at our institution. The two ventilators were compared for safety, reliability, ease of use, and efficacy in maintaining gas exchange. On VCV, end points of therapy were: fractional concentration of oxygen in the inspired gas (FIo,) 5 0.40; arterial oxygen pressure (Pa03 2 70 mm Hg; cardiac index (CI) 3.5 Umin/sq m; and spontaneous respiratory rate 5 eight breaths per minute. On HFJV, end points were: FIo, 5 0.45; arterial oxygen saturation r 0.90; and CI r 3.5 Wminlsq m. Spontaneous ventilation and pulmonary venous admixture reduction were the goals on VCV, with oxygen transport the goal on HFJV, Total duration of use of the ventilators was approximately 800 days with both types of devices; there were no technical failures, and the incidence of barotrauma was less than 5 percent. The end point

T

he cornerstone of therapy for acute respiratory failure is positive-pressure ventilation with positive end-expiratory pressure (PEEP).' Although disagreeamong authors on the ideal levels of ment still PEEP,2.3 scarcely an intensive care unit (ICU) exists wher; this treatment is not administered' to the vast majority of patients with respiratory f a i l ~ r e . ~ The therapeutic benefits of PEEP are achieved by increasing functional residual capacity (FRC). Collamed alveoli are reex~anded,&d alveoli with low cr%ical closing pressurk are stabilized. Matching of ventilation and perfusion is improved, while the enlarged volume presumably places the lung in a more favorable wsition on its ~ressure-volumecurve, reducing expenditures of energy required for tidal exchange~.~ Therapy with PEEP provides a unified approach to the management of respiratory failure; the increased pulmonary volume simultaneously improves arterial *From Memorial Sloan-Kettering Cancer Center, New York. tclinical Chief, Department of Critical Care. $Chairman, Department of Critical Care. $Technical Director, Department of Respiratory Therapy. enior Biomedical Engineer, Department of Medical Physics. \sResearch Assistant, De artment of Critical Care. **Assistant Attending ~K~sician, Department of Critical Care. Manuscript received July 26, 1982; revision accepted May 19. Reprint requests: D,:Carlon, 1275 York Avenue, New York City 10021

of mechanical ventilation was reached by a significantly higher percentage of the patients randomized to HFJV. Patients who failed to reach the therapeutic goal within 24 hours were crossed over to the other form of support. Those crossed from VCV to HFJV improved more rapidly and in greater number than those crossed from HFJV to VCV. When survival and total duration of stay in the intensive care unit were considered, there was no difference between VCV and HFJV. Considering data on gas exchange, VCV provided a higher WO, at equivalent positive end-respiratory pressure than HFJV. Alveolar ventilation was slightly better on HFJV. Differences were statistically but not clinically significant. On HFJV, oxygenation and ventilation were maintained with lower peak inspiratory pressures and smaller tidal volumes than those required for VCV. This investigation proves that HFJV is a safe and reliable method to provide mechanical support which does not, at this time, offer obvious benefits over VCV.

oxygenation and alveolar ventilation. Despite the unquestionable efficacy of therapy with PEEP, it has been known for many years that the condition of some patients with respiratory failure, perhaps as many as 20 percent, does not improve and may even deteriorate while their FRC increase^.^,' Patients with barotrauma, tracheoesophageal and bronchopleural fistulas, pathologically increased pulmonary volume, and diffuse pulmonary fibrosis are also among those whose condition may fail to respond to therapy with PEEP8 In respiratory failure, convective gas distribution within the lungs is often abnormal. Delivering larger tidal volumes, which must rely on convection for effective distribution, may not represent the best method to overcome the underlying pathologic abnormality.' Over the last 15 years, numerous researchers have indicated that adequate alveolar ventilation and arterial oxygenation can be maintained with ventilators which have negligible internal compressible volume and deliver tidal volumes substantially smaller than those of conventional ventilators. These devices are very different from each other, but they all operate at respiratory rates which are higher, and sometimes substantially higher, than those accepted as "normal. CHEST 1 84 1 5 I NOMMBER, 1983

551

A detailed discussion of each type of high-frequency ventilator, of the differences between them, and of the physiologic implications and potential advantages of each one is outside the scope of this report; however, there is a clear need to investigate alternative methods of mechanical ventilation and to determine their efficacy when conventional ventilation with PEEP fails. Anecdotal reports and small clinical series have been published on the use of high-frequency jet ventilation (HFJV) in a variety of clinical condition^.^^ An extensive and objective evaluation of the usefulness and limits of HFJV is not available at this time. The present investigation compares HFJV to volume-cycled ventilation (VCV) in a large series of patients who were randomly assigned to either type of ventilatory support. The primary goal of this study was to obtain substantive evidence that a new method of ventilation could be effectively applied in supportive therapy for acute respiratory failure. To reach a valid conclusion, the following technical and clinical characteristics were considered: (1)At the present stage of its technical development, is HFJV a safe and reliable method of mechanical ventilatory support, which can be routinely used in an ICU without requiring extraordinary personnel support? (2) Is there a difference in morbidity and mortality between patients supported with the two methods ofventilation?(3)Is HFJV equal, superior, or inferior to VCV in providing alveolar ventilation and arterial oxygenation in patients requiring mechanical assistance? (4) Can specific patients or pathologic conditions be identified, where respiratory function more consistently improves with one type of ventilation than with the other? (5)In patients who are successfully weaned, is there a difference in the total duration of mechanical support with the two types of ventilators? (6)Can guidelines be defined to adjust the ventilator's settings on HFJV, so that desired changes in arterial oxygen pressure (PaOJ or arterial carbon dioxide tension (PaCOJ or both may be consistently and reliably obtained? All patients admitted to the ICU of our institution between July 1, 1981 and Nov 30, 1982 were considered as candidates for the study. Criteria for admission into the investigation were as follows: (1) body weight of 20 kg (44 Ib) or more; and (2) respiratory failure requiring mechanical ventilation, associated with (a)diffuse bilateral infiltrates with refractory hypoxemia following major surgical procedures or sepsis in nonimmunocompromised patients, (b)postoperative aspiration pneumonia, (c) sepsis in patients with systemic malignant neoplasms (lymphomas and leukemias) during the bone marrow suppression which follows chemotherapy and radiation therapy, (d) diffuse pulmonary injury after administration of chemotherapeutic agents with specific pulmonary toxic effects, or (e) diffuse pulmonary injury after bone marrow transplantation. Patients in groups a and b were considered immunocompetent; those in groups c, d, and e were considered immunocompromised. Patients who required less than 12 continuous hours of mechanical support were excluded from

the study. During the period of the study, 321 patients were eligible for the investigation. Twelve of them could not be randomized because consent was unavailable. Three hundred and nine patients were assigned to HFJV and VCV according to a table of random numbers; at the conclusion of the study, 157had received VCV, and 152had received HFJV. The HFJV was delivered with a ventilator operating with a solenoid valve, which has been extensively described elsewhere.lB The tidal volumes generated by the device when the injector cannula was placed in a 10-mm endotracheal tube and when gas source pressure was progressively increased are reported in Figure 1. This information is provided as a mean of standardizing our ventilator, in conformity with recommendations of a workshop on high-frequency ventilation, sponsored by the National Heart, Lung, and Blood Institute, the American Thoracic Society, and the American College of Chest Physicians (New Orleans, Dec 5-7, 1982). For all applications of HFJV, the respiratory rate was 100 breaths per minute, the inspiratory-expiratory ratio was 1:2, and the internal diameter of the injector cannula was 1.62 mm (14 gauge). The ventilator circuit included a demand valve (Bear Medical) to provide gases for entrainment and a threshold resistor in the expiratory limb to allow addition of PEEP: The demand valve has a peak flow of 100 Umin to sustain constant positive airway pressure. An ultrasonic spirometer (Bear Medical)was placed in the expiratory limb, distal to the threshold resistor, to measure total exhaled volume. The response time of the spirometer provides accurate measurement when expiration lasts more than 300 msec (Fig 2).

Driving Pressure (PSIG) FIGURE1. Changes in tidal volume (V +) when driving pressure was increased from 5 to 35 psig; 1.62-mm jet cannula was inserted into endotracheal tube with 10-mm diameter, which was mounted into 80-cm corrugated airway hose with diameter of 2.4 cm. Distal end of those was connected to pneumotachygraph (Hans Rudolph model 3700). Drop in pressure across pneumotachygraph was measured by 0-2 cm full-smle differential pressure transducer (Gaeltech model 8T-2). Output of pressure transducer was amplified to provide flow wave form, which was observed on oscilloscope to verify absence of turbulent flow. Flow wave form was electronically integrated to provide volume determinations. Measurements were obtained three times at each driving pressure; values were always within 1percent of each other. High-frequency Jet Ventllatbn (Carlon eta/)

1

Solenoid

Optional peep attachment Entrained gases were humidified with a heated Cascade humidifier placed between the demand valve and the inspiratory line. Jet gases were humidified by delivering 20 m h r of physiologic saline solution through a narrow cannula running parallel to and opening perpendicular to the n o a l e of the injector cannula (Fig 3). Approximately half of the patients studied were intubated with an endotracheal tube (Hi-Low; National Catheter Corp.). This tube has a port which opens 1cm above its distal end, 20 to 25 cm from the injector cannula. The tube was placed 2 to 3 cm above the carina. Airway pressure was measured from this port using a pressure transducer (Bentley) connected to an oscilloscope (Abbott Medical Electronics). Peak and end-expiratory pressures were obtained by planimetric measurements, and mean airway. pressure was deter. mined by electronic integration. Patency of the pressure-measuring port was verified by examination of the o s c i l l o ~ p i ctrace before each measurement. The VCV was delivered with a commercially available ventilator (Bear Medical BEAR 1). which provides digital display of exhaled tidal volume and manometric values of peak inspiratory pressure. During their stay in the ICU, patients were treated with standard types of therapy for all nonrespiratory aspects of their management. The electrocardiogram was continuously monitored in all patients. When indicated, blood pressure was monitored through an indwell-

FIGURE2. Schematic representation of jet ventilator circuit, highlighting inspiratory and expiratory limbs, entrainment port, and location of PEEP valve.

ing arterial line and pulmonary arterial pressure through a pulmonary arterial catheter. Vital signs were recorded at least every hour or more frequently if appropriate. Cardiac output was measured by thermodilution at least three times per day. Samples of arterial and mixed venous blood for determinations of blood gas levels were obtained from the indwelling arterial line and the distal port of the pulmonary arterid catheter. These measurements were made several times per day in stable patients and after each ventilator change in unstable patients. To normalize PaO. for different values for the fractional concentration of oxygen in the inspired gas (FIoJ, the alveolarlarterial oxygen .'~ values are 500 to 600 tension ratio was used ( P ~ O ~ F I O JNormal units. Derived hemodynamic, respiratory, and oxygen transport variables were obtained with appropriate calculations.18 The absence of uniformly accepted goals for management of patients on conventional VCV created some difficulty when we attempted to define the "standard" against which HFJV would be compared. The literature suggests that reexpansion of pulmonary volume through PEEP, for the dual purpose of improving arterial oxygenation and facilitating spontaneous ventilation, is the preferred method of mechanical support in a majority of institutions, so this approach was used as "standard" for VCV. High-frequency jet ventilation is a form of controlled ventilation,

FIGURE3. Diagram of jet injector cannula, demonstrating twn parallel lines, one carrying gases and other carrying fluids for humidification. CHEST I 84 1 5 I NOVEMBER, 1983

553

since spontaneous breathing is generally suppressed, even at normocapnia. Expansion of pulmonary volume is only necessary to restore life-supporting arterial oxygenation at nontoxic FIo,, not to maintain spontaneous breathing. This difference between VCV and HFJV justified the use of different end points of therapy for the two methods of ventilation. As a consequence of this decision, two additional issues were addressed by the present investigation: (1)Are there differences in morbiditv. ,, mortalitv. ,. or the duration of ventilatory support if patients breathe spontaneously while on mechanical ventilation, as compared to full mechanical support? (2) Is there a difference in morbiditv. mortality or the total duration of mechanical ventilation if the primary end point of therapy is reduction of pulmonary venous admixture as compared to oxygen availability, independent of venous admixture? On VCV, patients were initially placed on FIo, of 0.6 to 0.8, PEEP of 5 cm H,O, tidal volume of 12 mYkg, and rate of intermittent mandatory ventilation (IMV)of ten breaths per minute. The goals of therapy were: (1)reducing QspIQt s 15 percent; to this end, PEEP was increased in increments of 2 to 3 cm H,O; (2) decreasing rates of IMV to four to six breaths per minute as long as pH was 7.35 to 7.45 and spontaneous respiratory rate s 35 breaths per minute; (3)reducing FIo, 5 0.40 as long as PaO, was 2 70 mm Hg; (4) maintaining cardiac index z 3.5 Uminlsqm using volume loading or inotropic agents (or both), as needed; and (5) maintaining hemoglobin levels 2 13 g1100 ml, with transfusions of packed red blood cells (RBCs), as needed. When PEEP had been decreased to 5 cm H,O and IMV to two breaths per minute, a trial of spontaneous ventilation was initiated. If this was successful, the patient was extubated. Patients randomized to HFJV were initially placed on FIo, of 0.6 to 0.8 and PEEP of 5 cm H,O. Driving pressure was adjusted to provide a tidal volume of 4 mYkg, which had maintained PaCO, within the desired range in previous patients." The goals of therapy were: (1) maintaining arterial oxygen saturation (SaOe) r 0.90, using FIo, S or 0.45 and increasing PEEP by increments of2 to 3 cm H,O, as needed; (2) maintaining PaCO, at 35 to 40 mm Hg, manipulating driving pressure as needed; (3) maintaining cardiac index of 3.5 Uminlsqm using volume loading or inotropic agents (or both), as needed; and (4) maintaining hemoglobin levels r 13 g/100 ml using transfusions of packed RBCs as needed. Spontaneous breathing was not encouraged, although no specific attempts to suppress it were made. Allowing considerable variations between patients and in the same individual at different times, spontaneous breathing was uncommon on HFJV. In patients whose condition improved, weaning was accomplished by decreasing PEEP to 5 cm H,O and decreasing driving pressure until continuous spontaneous breathing was detected. Trials of spontaneous ventilation were then initiated, and the patient was extubated upon their success. For ethical reasons, the study's design included a crossover alternative. Patients who had not reached the therapeutic goal within 2.4 hours on the ventilator to which they had been initially randomized were crossed over to the other support method. In these patients, 24 additional hours were allowed to reach the end point of therapy with the new ventilator. In case of failure, patients were maintained on the support method which had given the best PaOdFIo, ratio. The study had been approved by the investigational review board of our institution. Written informed consent was required for HFJV. Since institution of mechanical ventilation is often an emereencv , procedure, it was not always possible to obtain timely consent to immediately begin support with HFJV. In such cases, patients were initially placed on VCV and switched to HFJV as soon as consent was obtained. If this required more than three hours, as was the case in 12 instances, the patient was excluded from the study. The design of the study yielded two groups and four subgroups of patients, as shown in the following tabulation giving numbers of

Table 1-Utilizntion qf the Ventilators in the Study Data

HFJV

VCV

No. of ventilators used Total use, days* Longest use, dayst Average use, days* Failure of equipment No. with bamtrauma (percent)

6 770 14.8 2.5 None 4 (3)

9 854 23.6 2.5 None 6 (4)

*Sum of all periods of mechanical ventilation of all patients. t h n g e s t continuous application, without any interruption, of each type of ventilator. $Mean duration of each period of continuous use of ventilator. patients (number within parentheses are percents of group): Croup 1 (randomized to VCV) Subgroup 1A (end point reached in 24 hr) Subgroup 1B (end point not reached in 24 hr) Group 2 (randomized to HFJV) Subgroup 2A (end point reached in 24 hr) Subgroup 2B (end point not reached in 24 hr)

157 119 (76) 38 (24) 152 133 (88) 19 (13)

A significantly higher number of patients reached the end point of therapy on HFJV than on VCV (subgroup 1.4 vs 2A, xP=7.03; p c0.05).

Six different high-frequency ventilators and nine different volume-cycled ventilators were used, with no equipment failures and minimal iatrogenic complications (Table 1). Patients were evenly distributed between HFJV and VCV, considering the whole cohort and also on the basis of sex and underlying disease (Table 2). The rate of survival was identical with the two types of ventilation and so was the total duration of ventilation in the patients who survived and in those who died (Table 2). The end point of mechanical ventilation was reached more consistently in patients on HFJV (Table 3), at the same average level of PEEP as those on VCV. The time necessary to reach the end point after initiation of mechanical ventilation was longer on HFJV than on VCV. In the majority of cases, the immediate cause of death was not a failure of the mechanical ventilator to maintain life-supporting gas exchange. When blood gas levels were measured within two hours of death, values fbr PaO, at or above the end point of mechanical ventilation were observed in 55 (71 percent) of the patients on HFJV and in 38 (60 percent) of those on VCV. The difference was not statistically significant (lible 3). The majority of patients who were crossed over from one type of ventilation to the other had improvement of their blood gas levels, reaching the selected end point of therapy within 0.45 to 0.85 days; however, very few patients ultimately survived ('hble 4). There were no differences between the two types of ventilation. When changes in respiratory or hemodynamic varia-

Table 2 - D b h t h Croup and Ventilator*

qfpwisnb butwean MF]V

Died

Survived No.

and VCV

Percent ICU, Stayindayst

No.

'IbcP1

Percent ICU, Stay in dayst

No.

, ICU, Stay in dayst

Total population (307) VCV HFTV Immunoincompetent (213) VCV HFTV Immunocompetent (96) VCV HFTV Male patients (175) VCV HFJV Female patients (134) VCV HFJV *For each group, total number of patients is shown in parentheses. ?Mean SE.

+

bles registered in the first 60 minutes after a change in ventilator were analyzed, no d S r e n c e was observed between patients crossed over from HFJV to VCV and those crossed over from VCV to HFJV @able 4). Patients randomized to HFJV, but initially placed on VCV while consent was obtained, experienced a decrease of the PaOdFIo, ratio in the first 30 minutes after initiation of HFJV; however, at 60 minutes, PaOdFIo, was substantially higher than the last value recorded on VCV (Table 4). Eighty-eight patients developed respiratory Eailure while in the ICU. Management of their ventilatory support was therefore completely controlled by the investigatingteam from its inception, including immediate assignment to the appropriate ventilatory method. Mortality in this group was no different than in the total population of patients @able 5). Table 3-Poticnts Who Reached End Point of Therapy d h i n 24 Hours on E d Qp qf Mechankul SupporP Data Total patients No. reaching end point (percent) Time to reach end point, dayst Baseline PEEP, cm HIOO Increment in PEEP, cm H1qI Data obtained within 2 hr of death No. above end point (percent)l No. below end point (percent)?

HFW

VCV

152

157

133 (88)

114 (76)

0.34+0.03$

0.25ko.02

Analysis of data on hemodynamic and respiratory variables does not demonstrate clinically relevant difkrences between HFJV and VCV in patients who survived and in those who died @able 6). 'Iidal volume was substantially lower on HFJV than on VCV and so was peak inspiratory pressure @able 6). On HFJV, driving pressure was higher in the patients who died, reflecting the requirement for larger tidal volume to maintain the desired PaCO, (Table 6). High-frequencyjet ventilation, at its present stage of

Table IccmmwerFon, One Qp qf Vsntilotor to the

w

FromVCV toHFJV

Tdpi patients crossed over No. surviving (percent) No. reaching end point (percent) T i e to reach end point, dayst Baseline PEEP, cm H,OS Change in PEEP, cm H,OO Immediate changes in oxygenation, F k O ~ o units , Baseline

6.620.6

5.0k0.4

1.9320.17

1.60k0.13

Change in 30 min

55 (71) 38 (60)

22 (29) 25 (40)

Change at 60 min

*Table values are means + SE unless otherwise stated. tTime from initiation of mechanical ventilatiotl. SP < 0.05. OLevel of PEEP observed when patients were first placed on mechanical ventilation. IlIncrease in PEEP required to reach end point. lAbove or below end point of therapy ~ccordingto arterial blood gas levels.

From HFJV toVCV

Data

*?8ble values are means 2 SE unless otherwise stated. tltme to reach end point &r crossaver. $Last PEEP recorded &re crosover. OMfference between basebe and level at which end point of therapy was reached. IlData within pnrentheses are from 81patients randomized to HFJV but initially p l e d on VCV h r up to 3 hr. 1 ' P < 0.05. CHEST 1 84 1 5 1 NOMMBER. 19a7

Table 5--Pa*

Who D m l o ~ e d

Fadun in

ICU*

Data

vCv

No. who survived (percent) No. who died (percent) Total patients

22 (37) 38 (63) 60

H F ~ l1 (39) 17 (61) 28

*Data from 88 patients who were immediately placed on ventilator to which they had been randomized.

development, can be routinely used for respiratory support in an ICU with a normal corriplement of nursing and respiratory therapy st&. The ventilators were used over a period of 1%years, during the time approxinlately 15new registered nurses were inducted into our ICU. High-frequency jet ventilation had been integrated into their orientation schedule and did not require prolongation of the training period. In almost 800 days of continuous use, not a single component of any of the jet ventilators failed. This result is not surprising, since the device only has one moving part, the solenoid valve, which is a well tested piece of equipment with a life expectancy of over 100 million cycles. Our present policy requires chanwg the solenoid valve eveiy three months, after an average use of 10 million cycles. The cost of replacing the solenoid valve is negligible. The incidence of barotrauma was less than 5 percent with both types of ventilators, an incidence which is acceptable according to most investigator^.'^.^ This obseivation supports the hypothesis that barotrauma is related to changes of parenchymal structure, rather than to the mechanical device used. Evaluation of the efficacy of HFJV in respiratory failure is a complex issue. Variables which affect ultimate survival are many, and gas exchange (PaO,

and PaCOJ is only one of them. In the population of patients studied, the relative uniformity of underlying pathologic abnormalities offered some statistical advantages but also biased the study with the incidence of complications associated with these diseases. The expected mortality of immunocompromised patients in respiratory failure is approximately 80 percent.2'.e2 Uncontrollable sepsis, hemorrhages, and hemodynamic instability are largely responsible for this rate. It is unrealistic ti expect&& a new mechanical device may substantiallyimprove those complications. On the other hand, adding any negative component to the support of very unstable patients can only have catastrophic consequences. The data presented indicate the substantial equivalence of VCV and HFW, when the duration of hospitalization and ultimate survival were the variables considered. This applies to the entire population studied and to each subgrbup (immunocompetent and immunocompromised patients, patients who developed respiratory failure in the ICU). There is no universally accepted method of support for patients with respiratory failure; however, it is possible to broadly divide existing philosophies of ventilatory management into two groups. Some researchers and cliniciaris maintain that reduction of shunt fraction and reexpansion of pulmonary volume must have absolute priority, even if very high levels of PEEP must be used and FRC expanded to 150 to 200 percent of predicted.== Other investigators suggest that oxygen transport should be the primary end point of therapy, and PEEP should be added only to reduce FIo, below toxic levels.ppes The role of spontaneous breathing in the management of respiratory &lure has not been satisfactorily

Table &Respiratory and Hemodyamic Variables and Raw Values ofVentilotor-Specific Variables* Survived VCV Variable Respiratory and hemodynamic PaO, mm Hg PaOJFIo, mm Hg PaCO,, mm Hg PEEP, cm H,O Mean systemic blood pressure, mm Hg Mean pulmonary arterial pressure, mm Hg CI, Wmintsq rn Qsp/% percent Ventilatordependent lldal volume, mYkg Peak inspiratory pressure, cm H,O Driving pressure, psig Respiratory rate, breaths per min

No.t

Value

963 963 963 909

296

94.4k1.1 237.622.7 42.320.3 5.720.1 88.620.5 23.3 k 0.2 4.2k0.1 13.021.0

829 789

11.220.1 37.220.4

854

8.9k0.2

982

642 282

...

...

HFJV

.

Died I

VCV

HFTV

Value

No. t

Value

No.t

Value

984 927 1,003 886 380 340

82.620.9 197.4k2.4 37.7k0.2 7.620.1 86.220.5 25.02 0.2 3.720.1 14.021.0

2,090 2,090 2,090 2,021 2,132 1,678 729 700

80.720.6 192.6k1.6 43.620.2 8.320.1 82.620.3 S . 4 k0.2 4.3k0.1 15.0k1.0

3,010 3,010 3,010 2,864 3,042 2,774 1,245 1,081

73.7k0.5 162.321.2 41.220.2 10.9-CO.1 83.5-CO.3 28.82 0.1 3.920.1 20.021.0

769 403 838

3.520.1 12.320.3 13.4k0.1

1,719 1,639

12.120.1 46.020.3

2,309 1,201 2,554

4.2k0.1 18.2e0.3 16.920.1

1,895

11.5k0.1

No.t 986

984

...

...

...

...

...

...

*Table values are means 2 SE unless otherwise noted. +Total no. of observations.

556

Highfrequency Jet Ventilation (Wonet d)

defined. Modern ventilators are designed to allow spontaneous breathing at high flow rates without excessive inspiratory eftbrt; however, the advantages of IMV over controlled ventilation are not universally recognized.% The two methods of ventilation used in the present study appeared ideally suited to address some of these controversial issues. Spontaneousventilation is usually suppressed by HFJV,30and ainvay pressure excursions around mean values are reduced. High peak inspiratory pressures are not generated, so that alveoli whose critical opening pressure is substantially higher than mean airway pressure may remain collapsed, increasing pulmonary venous admixture. As we administered it, HFJV can therefbre be considered a model for controlled ventilation at low pulmonary volumes, with oxygen transport as the primary end pint of therapy. Volume-cycled ventilation with IMV generates higher inspiratory pressures, which should provide additional alveolar recruitment and larger pulmonary volumes. The clinical application of VCV in our population of patients was therefore a representative model of the ventilatory strategy based on shunt reduction and compliance improvement. The end points of mechanical support are admittedly empiric fbr both methods of ventilation; however, they are based on two widely accepted principles of management, both with solid roots in the existing literature on acute respiratory failure. The present study indicates that the selected end point of mechanical ventilation is reached more frequently on HFJV than on VCV. Also, patients crossed over from VCV to HFJV reached the therapeutic end point more rapidly and in greater numbers than those crossed over from HFJV to VCV. We recognize the need to exercise considerable caution when interpreting these observations. Therapeutic goals were less demanding for HFJV than VCV, but this had no negative influence on survival, complication rates, or the duration of mechanical support. In the evaluation of any mechanical device, its ability to provide reliable results must be considered an important feature. In this respect, HFJV proved successful in maintaining arterial oxygenation and alveolar ventilation within limits compatible with safe management of respiratory failure and consistent with the goals selected before initiation of the study. From this restrictive point of view, HFJV must be judged at least as safe and effective as conventional ventilation. The events following crossover from one type of ventilation to the other provide material for additional considerations. Most patients with failure on one type of ventilation had an immediate and significant improvement of their oxygenation on the other method of support. Among patients with the most severe forms of respiratory failure, two different categories may thus

be envisioned.

Patients in the first category require elevated peak inspiratory pressures and large tidal volumes to maintain alveolar stability and arterial oxygenation; patients in the second category, more numerous in our study, benefit from smaller tidal volumes, delivered at constant airway pressure and with a distributive method not based on convection alone. Some researchers have associated infrequent large positive-pressure breaths with high-frequency ventilation, reporting satisfactory r e s ~ l t s . Such ~ ' a technique may combine the alveolar stabilizing effect of high peak inspiratory pressure with the reduction of convective gas distribution abnormalities of HFJV. Data from the subgroup of 81 patients randomized to HFJV but supported on VCV in the first few hours suggest that in the majority of cases, VCV and HFJV are equivalent. Oxygenation and ventilation were as effective with large or small tidal volumes and high or low peak inspiratory pressures. Immediately upon initiation of HFJV, those patients experienced a deterioration of their arterial oxygenation; however, within 60 minutes the PaOdFIo, had risen substantiallyabove baseline, although PEEP, on the average, had been increased by 1.2 cm H,O only. It is probable that the immediate effect of terminating positive-pressure ventilation was a collapse of alveolar units with high critical opening pressure. This adverse development rapidly reversed, as evidenced by the substantial improvement in oxygenation observed within 30 to 60 minutes after a change in ventilator. Although no objective proof is available, a reasonable explanation is that intra-alveolar gas distribution improved and perhaps pulmonary volume increased through hysteresis. These observations were obtained so close to the time of the change of ventilator that they could not have been influenced by improvement of the underlying pulmonary pathologic abnormalities. Analysis of data collected confirms that VCV can provide a better oxygenation than HFJV at equivalent or slightly lower PEEP, but substantially higher peak inspiratory pressure. Clearance of carbon dioxide is adequate with both types of ventilator; although observed values for PaCO, were lower on HFJV, the difference was not clinically significant. Mechanical minute ventilation was 30 percent higher on VCV and 20 percent on HFJV in patients who died, as compared to those who improved. This observation suggests that in the deteriorating lung, convective flow must be increased with both types of mechanical support, although less with HFJV. Hemodynamic function was not influenced by the ventilatory method. Formulation of guidelines for ventilatory management on HFJV is also possible on the basis of this CHEST I 84 1 5 1 NOVEMBER, 1983

557

study. A relationship between exhaled tidal volume and PaCO, was established; tidal volume of 3.5 to 4.5 mlkg provides a PaCO, of 35 to 45 mm Hg in 99 percent of the patients. A relationship between exhaled tidal volume and driving pressure was also established. In the range of 10 to 25 pounds per square inch gauge, exhaled tidal volume (the sum of jet flow and entrained gases) was linear (150 to 300 ml) and approximately twice as high as bench-mark values (Fig 1). On HFJV, arterial oxygenation improves in most patients with the addition of PEEP, which raises mean airway pressure. In this respect, HFJV is no different than conventional ventilation. We can offer no advice on the advantages of different respiratory rates, injector cannula sizes, or inspiratoryexpiratory ratio. In previous experiments in animals, no settings provided more favorable results than those used in this study.=" In conclusion, HFJV is a safe, predictable, reliable, and effective method of mechanical ventilation which has, however, no obvious advantage over VCV in the population studied. At comparable levels of PEEP, HFJV provides better alveolar ventilation but poorer arterial oxygenation than VCV. This result is obtained at lower peak airway pressure; this may be beneficial in patients who rapidly develop interstitial fibrosis when subjected to high distending alveolar pressure.% Intermittent mandatory ventilation or controlled ventilation as support methods and shunt reduction or oxygen transport as end points of therapy are comparable when mortality or duration of ventilation are considered. No data in the present study allow us to recommend HFJV as the preferred method of ventilation in the types of respiratory failure investigated. Similarly, no data have emerged to advise against the use of HFJV in acute respiratory failure, so that clinical trials can be continued without ethical prejudice. A new device is available to clinicians to treat a condition which still carries a poor prognosis. Greater clinical experience may indicate whether specific indications for or contraindicationsto the use of HFJV in respiratory failure exist. ACKNOWLEDGMENT: We thank Ms. Clara Ducharn for expert technical assistance in typing this manuscript. 1 Weissman IM, Rinaldo JE, Rogers RM. Positive endexpiratory pressure in adult respiratory failure. N Engl J Med 1982; 307:1381 2 Suter PM, Fairley HB, Isenberg MD. Optimum endexpiratory pressure in patients with acute pulmonaty failure. N Engl J Med 1975;292:2&4 3 Gallagher TJ,Civetta JM, Kirby RR. Terminology update: optimal PEEP. Crit Care Med 1978;6:323 4 Hudson LD, ed. Adult respiratory distress syndrome. Semin

Respir Med 1981;299

5 Downs JB, Klein E F Jr, Desautels D, Modell JH, Kirby RR.

Intermittent mandatory ventilation: a new approach to weaning patients from mechanical ventilators. Chest 1973;64:331 6 Powers SR, Mannal R, Neclerio M, English M, Marr C, Leather R, et al. Physiologic consequences of positive endexpiratory pressure (PEEP) ventilation. Ann Surg 1973;178:265 7 Springe RR, Steven PM. The influence of PEEP on survival of patients in respiratory failure. Am J Med 1979;66:1978 8 Zapol WM, Tralstad RL, Coffey JW, Tsai I, Salvador RA. Pulmonary fibrosis in severe acute respiratory failure. Am Rev Respir Dis 1979;119:547 9 Kolton M, Cattran CB, Kent G, Volgyesi G, Froese AB, Bryan AC. Oxygenation during high-frequency ventilation compared with conventional mechanical ventilation in hm models of lung injury. Anesth Analg 1982;61323 10 Sjostrand U. High frequency positive pressure ventilation (HFPPV): a review. Crit Care Med 1980;8:345 11 Klain M, Smith RB. High frequency percutaneous transtracheal jet ventilation. Crit Care Med 1977;5:280 12 Butler WJ, kohn DJ, Bryan AC, Proese AB. Ventilation of highfrequency oscillation in humans. Anesth Analg 1980;59:577 13 Carlon GC, RayC Jr, Klain M,McCormack PM. High frequency positive pressure ventilation in management of a patient with bronchopleural fistula. Anesthesiology 1980;52:160 14 Carlon CG, Ray C Jr, Pierri MK, Groeger J, Howland WS. High Frequency jet ventilation. Chest 1982;81:350 15 Carlon GC, Khan RC, Howhod WS, Ray C Jr. 'Iiunbull AD. Clinical experiencewith high fmquency jet ventilation. Crit Care Med 1981;9:l 16 Carlon GC,Miodownik S, Ray C Jr, Kahn RC. Technical aspects and clinical implications of high frequency jet ventilation with a solenoid valve. Crit Care Med 1981;9:47 17 Keighley GR. The arteriaValveolar oxygen ratio: an index of gas exchange applicable to varying inspired oxygen concentrations. Am Rev Respir Dis 1974;109:142 18 Cohn JD, Engler PE. Shunt effect ofcarboxyhemoglobin. Crit Care Med 1979;7:s 19 Kirby RR. Ventilatory support and pulmonary barotrauma. Anesthesiology 1979;50:1981 20 Lewis FR,Blaistell FW,Schlobohm RM. Incidence and outcome of posttraumatic respiratory failure. Arch Surg 1977;112:436 21 Hwlett RF, Wilson AE Adult respiratory distress syndrome (ARDS) bflowing aggressive management of extensive acute lymphoblastic leukemia Cancer 1977;39:2422 22 Carlon CC. Respiratory failure in cancer pafients. Curr Pmbl Cancer 1978;4:3 23 Gallagher TS,Civetta JM. Coal directed therapy of respiratory failure. Anesth Analg 1980;59:831 24 Kirby RR, Downs JB, Civetta JM, Modell JH, Damemiller FJ, Klein EF, et al. High level positive endexpiratory pressure (PEEP) in acute respiratory insdiciency. Chest 1975;67:156 25 Rose DM, Downs JB, Meenan TJ.Temporal response of funo tional residual capacity and oxygen tension to changes in positive endexpiratory pressure. Crit Care Med 1981;9:79 26 Pontoppidan H, W n B, Lowenstein E. Acute respiratory Wure in the adult. N Engl J Med 287:690,1972 27 Gong H Jr. Positive pressure ventilation in the adult respiratory distress syndrome. Clin Chest Med 1982;3:69 28 Qvist J, Pontoppidan H, Wilson RS, Lowenstein E, Caves MB. Hemodynamic responses to mechanical ventilation with PEEP: the effect of hypervolemia. Anesthesiology 1975;42:G 29 Petty TL.Intermittent mandatory ventilation reconsidered. Crit Care Med 1981;9:620 30 Jonzon A. Phrenic and vagal nerve activities during spontaneous ventilation and positive pressure ventilation. Acta Anaesthesiol Scand (suppl) 1977;64:29

31 El-Baz N, Faber Penfield L, Doolan A. Combined high-frequency ventilation for management of respiratory failure: a new technique. Anesth Analg 1983; 62:39 32 Carlon GC, Ray C, Griffin J, Miodownik S, Groeger JS. lidal volume and airway pressure on high frequency jet ventilation. Crit Care Med (in press) 33 Carlon CC, Ray C, Coetz WS, Groeger J. High frequency jet

ventilation in respiratory failure: influence of driving pressure and cannula size. Crit Care Med 1981; 9:159 34 Webb HH, lierney DE Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressure: protection by positive end-expiratory pressure. Am Rev Respir Dis 1974; 110:556

Percutaneous Transluminal Coronary Angioplasty Training Course The Texas Heart Institute will sponsor this course to be held January 26-28 at the Institute (Houston). For information, contact the Medical Director's OI%ce, Texas Heart Institute, PO Box 20269, Houston, Texas 77225 (713:791-2157).

Keystone Summit on Allergy, Immunology, Pulmonology and ENT The National Jewish Hospital and Research Centerrnational Asthma Center and National Jewish Hospital Continuing Medical Education Committee will present this course January 29February 2 at Keystone, Colorado. For information, contact: Helga Cole, NJWNAC, 3800 East Colfax Avenue, Denver 80206.

CHEST 1 84 1 5 I NOVEMBER. 1983

559