Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous...

153
VU Research Portal Preservation of spontaneous breathing during high-frequency oscillatory ventilation van Heerde, M. 2011 document version Publisher's PDF, also known as Version of record Link to publication in VU Research Portal citation for published version (APA) van Heerde, M. (2011). Preservation of spontaneous breathing during high-frequency oscillatory ventilation. General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ? Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. E-mail address: [email protected] Download date: 07. Feb. 2021

Transcript of Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous...

Page 1: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

VU Research Portal

Preservation of spontaneous breathing during high-frequency oscillatory ventilation

van Heerde, M.

2011

document versionPublisher's PDF, also known as Version of record

Link to publication in VU Research Portal

citation for published version (APA)van Heerde, M. (2011). Preservation of spontaneous breathing during high-frequency oscillatory ventilation.

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

E-mail address:[email protected]

Download date: 07. Feb. 2021

Page 2: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Preservation of spontaneous breathing during high-frequency oscillatory ventilation

Marc van Heerde

Page 3: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

The research presented in this thesis was performed at the Department of Pediatric Intensive Care, VU University Medical Center, Amsterdam, the Netherlands, the Faculty of Biomedical Engineering of the Czech Technical University in Prague, Kladno, Czech Republic and the Central Laboratory Animal Research Facility, University Medical Center Utrecht, Utrecht, the Netherlands. The research was embedded within the Institute for Cardiovascular Research of the VU University Medical Center (ICaR-VU). This project was financially supported by VIASYS Healthcare (now CareFusion) and by the Ministry of Education, Youth and Sports of the Czech Republic, project number MSM 6840770012. Cover Buurman & Buurman, © Patmat, Prague, Czech Republic, 2011 Copyright © Marc van Heerde, Amsterdam, The Netherlands, 2011

ISBN 978 90 865 9576 1 Printed Ipskamp Drukkers

The Publication of this thesis was financially supported by Stichting Kind & Afweer, Stichting Researchfonds Kindergeneeskunde and CareFusion.

Page 4: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

VRIJE UNIVERSTITEIT

Preservation of spontaneous breathing during high-frequency oscillatory ventilation

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad Doctor aan

de Vrije Universiteit Amsterdam, op gezag van de rector magnificus

prof.dr. L.M. Bouter, in het openbaar te verdedigen

ten overstaan van de promotiecommissie van de faculteit der Geneeskunde

op vrijdag 2 december 2011 om 11.45 uur in de aula van de universiteit,

De Boelelaan 1105

door

Marc van Heerde

geboren te Zwolle

Page 5: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

promotor: prof.dr. J.J. Roord copromotor: dr. D.G. Markhorst

Page 6: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Contents

Chapter 1 General introduction and outline of this thesis 7

Chapter 2 Imposed work of breathing during high-frequency oscillatory ventilation: a bench study Critical Care 2006

21

Chapter 3 Design and control of a demand flow system for a high-frequency oscillatory ventilator IEEE Transactions on Biomedical Engineering 2011

33

Chapter 4 Unloading work of breathing during high-frequency oscillatory ventilation: a bench study Critical Care 2006

51

Chapter 5 Demand flow facilitates spontaneous breathing during high-frequency oscillatory ventilation in a pig model Critical Care Medicine 2009

62

Chapter 6 Effect of spontaneous breathing during high-frequency oscillatory ventilation on regional lung characteristics in experimental lung injury Acta Anaesthesiologica Scandinavica 2010

77

Chapter 7 Reflections on pediatric high-frequency oscillatory ventilation from a physiologic perspective Respiratory Care in press

91

Chapter 8 General discussion 105

Chapter 9 Summary 117

Samenvatting 123

References 131

Authors’ Affiliations 146

Dankwoord 147

Curriculum vitae 151

Page 7: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

The Respiratory Dead Space

As several of the recent investigators in this field have evidently left out of account the peculiarities of the movements of liquids and gases in tubes, we may begin the report of our data with a few simple experiments on this topic. They illustrate why it is that with a dead space of 150 cc. the first 150 cc. expired (or even the first 50 cc.) are not free from a considerable admixture of pulmonary air. They afford the reason why a dog during heat polypnoea may have a tidal volume considerably less than the volume of the dead space. They suggest that, during rapid shallow breathing in man, what we may call the physiological dead space is a much smaller volume than the anatomical dead space of Loewy’s plaster cast. There may easily be a gaseous exchange sufficient to support life even when the tidal volume is considerably less than the dead space. Yandell Henderson, FP Chillingworth and JL Whitney in The American Journal of Physiology Volume 38, July 1, 1915.

Page 8: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

General introduction & outline of this thesis

Page 9: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 1

8

Background

The simple automatic act of breathing is essential to life, but may be interrupted in a variety of ways. Cessation of this act of breathing is not necessarily fatal if there is some means available to produce artificial respiration. The first reference to artificial respiration is in Egyptian mythology, about 1500 B.C., where by one account Isis resurrected Osiris with ‘the breath of life’.1 The history of artificial ventilation is usually regarded as beginning with the 16th century anatomist Andreas Vesalius. In the renaissance progress was made in physiology of respiration and Vesalius was the first to describe, in 1543, a technique to mechanically ventilate an animal while examining its thoracic contents.2 From the 16th century onwards artificial ventilation further developed. Early in the 19th century already, doubts were raised concerning the safety of positive pressure ventilation. The proof was delivered by Leroy in 1828. In an experiment he tried to recover animals which had been drowned. The pulmonary emphysema that was found was blamed on the bellows used for ventilating the animals.3 The dangers of positive pressure ventilation led to division of opinion on these matters. In England, the Royal Humane Society abandoned the use of bellows in their resuscitation scheme at that time. To the opposite the Scottish ‘Edinburgh Almanack’ of 1834 advised the use of bellows during resuscitation, but warned to be gentle while inflating the lungs.4 The modern era of mechanical ventilation and intensive care began during the polio pandemic in the 20th century. In 1953 Lassen published a classic paper on the use of mechanical ventilation in patients with paralytic polio.5 Just by applying mechanical ventilation, he described a dramatic and sudden drop in mortality from over 80% to around 40%. It led to wards where paralytic polio patients were ventilated. It was the start of modern intensive care units. Polio has been eradicated. Now, the conditions that we are concerned about the most in terms of mechanical ventilation are pulmonary diseases like the acute respiratory distress syndrome (ARDS). Since the 1950s numerous animal and clinical studies have shown that mechanical ventilation itself can initiate as well as exacerbate lung injury.6 This phenomenon is known as ventilator-induced lung injury (VILI). Once the mechanisms responsible for VILI were elucidated, ventilator strategies were sought that could protect the already injured lung from additional harm. High-frequency oscillatory (HFO) ventilation is one of these strategies.

Page 10: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General introduction and outline

9

Acute respiratory distress syndrome In 1967 Ashbaugh was the first to describe ARDS in 12 adults.7 The respiratory distress syndrome in the 12 patients was manifested by acute onset of tachypnea, hypoxemia, and loss of respiratory system compliance. The etiology of ARDS varied, the response of the lungs was however similar in all 12 patients. The clinical and pathological features closely resembled those seen in infants with respiratory distress syndrome. Ashbaugh and co-workers postulated the theoretical relationship of this syndrome to an alveolar surface active agent. They observed that positive end-expiratory pressure (PEEP) and corticosteroids appeared to have value in the treatment of the patients. ARDS, more than 40 years after its first description, still has significant mortality and morbidity. Mortality rates for adults range from 30 to 60%.8-10 In the pediatric age group mortality is usually less but also significant, up to 30%.11,12

Pathogenesis of acute respiratory distress syndrome

Endothelial and epithelial injury Efforts to understand acute lung injury (ALI) or the more severe ARDS, have been substantial. In a healthy lung the blood-gas barrier consists of the capillary endothelium and the alveolar epithelium.13 Within a decade after the first description of ARDS, clinical and experimental studies established the concept that increased permeability of the blood-gas barrier is the primary physiological abnormality in the early stages of ALI/ARDS. Endothelial injury and increased capillary permeability result in the formation of protein rich pulmonary edema.14-17 Injury to the alveolar epithelium has numerous consequences. First, the diffuse alveolar injury contributes to alveolar flooding and also impedes removal of edema fluids from the alveolar space. Second, damage to alveolar type II cells impedes surfactant production and turnover. Surfactant lowers surface tension and plays a key role in alveolar stability. Surfactant inactivation contributes to alveolar collapse.18,19 Finally, the inflammatory cascade is activated with release of numerous mediators. Activated alveolar macrophages secrete pro-inflammatory mediators to stimulate chemotaxis and activate neutrophils. Neutrophils migrate to the alveolar airspace and, in turn, release pro-inflammatory mediators and amplify the inflammatory response in ALI/ARDS. The diffuse alveolar damage results in the disturbance of oxygen uptake and poorly compliant lungs. The activated inflammatory response in the lungs not only causes

Page 11: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 1

10

injury to the lung itself. Translocation of mediators to the circulation can lead to the development of multisystem organ failure.6,19

Role of ventilator-induced lung injury The potential toxic effect of high fractions of inspired oxygen to the lungs were long recognized.20,21 In the last decades there is an expanding insight that the action of mechanical forces on the lung during mechanical ventilation itself can initiate or aggravate lung injury.6 To further improve lung protective ventilation strategies it is pivotal to understand the mechanisms responsible for VILI. An overview of seminal studies, which contributed to the understanding of VILI, is summarized below.

Overview of important animal studies on ventilator-induced lung injury In 1939 Macklin demonstrated that excessive alveolar distension produced air leaks. The application of high airway pressures seemed to cause the air leaks. The term ‘barotrauma’ was introduced.22 Greenfield, in 1964, showed in dogs, that large tidal volumes generated with peak inspiratory pressures (PIP) up to 36 cmH2O caused surfactant dysfunction.23 Another important observation in the early years of VILI research was that ventilation in the diseased lung is inhomogeneous.24 In 1970 Mead demonstrated that traction forces in an inhomogeneous aerated lung could exceed 140 cmH2O. These high forces were observed in lung tissue between an expanded alveolus adjacent to a collapsed alveolus.25 Webb and Tierney studied the effect of different ventilatory pressures, PIP and PEEP, on normal rat lungs.26 They showed that ventilating with high PIP, 30 to 45 cmH2O, resulted in severe lung injury and death within 35 minutes. They also demonstrated that PEEP protected against pulmonary edema caused by high peak pressures. Dreyfuss did an important observation in experiments with ventilated rats.27 Rats were ventilated with different strategies in order to differentiate the effects of tidal volume Vt and airway pressures on the lungs. In one strategy, for example, a thoracoabdominal binder limited Vt during ventilation with high PIP. They showed that high Vt, irrespective of airway pressure, produced lung injury. Lung injury did however not occur when the Vt was low, irrespective of the airway pressures. From that time on, focus of research moved from barotrauma to volutrauma. In fact the two terms are linked via the pressure volume relation of the lungs. An excessive transpulmonary pressure, the difference between the alveolar pressure and the pleural pressure, causes injurious overdistension, Figure 1.28,29

Page 12: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General introduction and outline

11

Among others, Muscedere demonstrated that repetitive opening and closing of distal lung units resulted in lung injury.30 The term atelectrauma was introduced. The experiments were performed in an ex vivo rat lung model. The occurrence of lung injury was studied in relation to a specific point on the inflation limb of the pressure volume curve of the lungs, the lower corner pressure (LCP). The LCP has usually been interpreted as the point where all previously collapsed alveoli are recruited during inflation. Muscedere showed that using PEEP below the LCP on the inflation limb of the pressure volume loop caused significant lung injury and impaired lung compliance compared to using PEEP above the LCP. At the present time it is known that recruitment occurs at pressures higher than the LCP, but this does not weaken the observations made by Muscedere.31 In the late 1980s another concept was introduced in unraveling the mechanisms responsible for VILI: biotrauma. There was a growing insight that not only mechanical forces produce VILI, also the inflammatory response contributes to it. Hamilton, in 1983, performed an experiment in surfactant depleted rabbits.32 HFO ventilation, using mean airway pressure mpaw of 15 cmH2O, Vt 1.5 ml/kg, frequency f 15 Hz, was compared to conventional ventilation, using a PIP of 25 cmH2O and a PEEP of 6 cmH2O. The HFO ventilated rats showed significantly better lung function and had fewer signs of histological lung injury. Histology of the lungs of

Figure 1 Transpulmonary pressure. Left panel: measurement of transpulmonary pressure under static conditions. The transmural pressure (ptm) is the difference between alveolar (palv) and pleural (ppl) pressure: ptm = palv – ppl. The esophageal pressure (pes) is a near approximation of the pleural pressure. At conditions of zero flow the pressure at the airway opening (pao) is equal to the alveolar pressure, then: ptm = pao – pes. Right panel: lung volume plotted in relation to transmural pressure. Note that the inflation and the deflation curves are not the same; this is called hysteresis.

Page 13: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 1

12

the conventional ventilated rabbits showed granulocyte infiltration in the alveoli and interstitium. The studies that followed elucidated the role of inflammatory cells and mediators in VILI. In 1998 it was hypothesized that by overspill of inflammatory mediators multiple system organ dysfunction could develop.33

Overview of important clinical studies on ventilator-induced lung injury Gattinoni and co-workers studied the computer tomographic (CT) appearance of lungs in mechanically ventilated patients suffering from ALI.34,35 The CT scans revealed that the lungs were affected very heterogeneously. The dorsal dependent lung areas were fluid filled and/or atelectatic. During the respiratory cycle hardly any change was seen in these parts of the lungs. The ventral nondependent lung zones were aerated and appeared to be the only part of the lungs that were ventilated during the respiratory cycle. It led to the insight that, in an inhomogeneous diseased lung, the applied Vt only was distributed to a small part of the lungs and in fact could produce significant regional overdistension. The small aerated part of the lung was named the baby lung. The concept of the baby lung, barotrauma, volutrauma etc. changed the mindset of clinicians how to ventilate a patient. Instead of aiming at normalization of arterial blood gas values using Vt’s of 10–15 ml/kg body weight, lungs were ventilated more gentle, with lower volumes and lower peak pressures, accepting a higher PaCO2 and somewhat diminished oxygenation. In 1990 Hickling reported the results of a study that used low Vt ventilation in 70 patients with severe ARDS.36 PIP was kept below 30–40 cmH2O resulting in an applied Vt of 4–7 ml/kg body weight. There was a 60% reduction in mortality compared to the mortality that was predicted based on a severity of disease classification system. Despite important limitation of this study it helped to change the approach of mechanical ventilation towards a protective strategy. Several prospective trials examined the effect of Vt on patient outcome.37-41 The ARDSNet trail, the largest multi-center prospective randomized trial including n = 861 ALI/ARDS patients, clearly showed a benefit of a low Vt strategy. The protective ventilation strategy resulted in reduced mortality, 31% vs. 39.8%, and more ventilator free days. The ARDSNet trial, in accordance with the earlier discussed animal studies, suggested that a lung protective ventilation strategy reduced both systemic inflammation and nonpulmonary organ failure. Another influential clinical study examining the effect of mechanical ventilation on the pulmonary and systemic inflammatory response was conducted by Ranieri and

Page 14: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General introduction and outline

13

co-workers.43 This study showed that a lung protective ventilation strategy reduced the pulmonary and systemic inflammatory cytokine response.

Mechanical ventilation

Lung protective mechanical ventilation A lung protective ventilation strategy aims at reducing potentially damaging mechanisms and factors causing VILI: oxygen toxicity, barotrauma, volutrauma, atelectrauma and biotrauma. Lachmann, in 1992, introduced the open lung concept.44 The open lung approach for lung protective ventilation was depicted by Froese, in 1997.42 She divided the pressure volume loop of the lungs in hazardous zones and a safe window. To better understand the optimal position of ventilation the original figure of Froese needs some modification, Figure 2. Regional atelectasis and overdistension are gradual phenomena in a diseased respiratory system. When recruiting the lungs on the inflation limb for example, some lung regions are already overdistended while other regions still are atelectatic.45 Therefore the safe window within the pressure volume loop is limited

Figure 2 Pressure volume loop of moderately diseased lungs, illustrating the hazardous zones. At low airway pressures the lung is atelectatic and derecruitment occurs (1). Increasing the airway pressure results in opening of collapsed lung areas. Simultaneously some lung zones already over distend (3). Further increase in airway pressure recruits almost all alveoli. At the highest airway pressure the total lung is overdistended (2). When, from that point, airway pressure is lowered, the overdistension gradually decreases. In the area designated as safe window the lung units are open and regional overdistension, atelectasis and derecruitment are minimized. Adapted from Froese (with permission).42 

Page 15: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 1

14

to the deflation limb of the curve. After a recruitment maneuver, PEEP and PIP are titrated to limit derecruitment and overdistension. The theoretical basis of different lung protective ventilation strategies is depicted in Figure 3. The management of ALI/ARDS is not only limited to ventilator settings. A wide variety of adjunctive and supportive therapies were studied. Prone positioning and inhaled nitric oxide both improve oxygenation, but do not improve overall clinical outcome.46,47 Extracorporeal membrane oxygenation (ECMO) has been used as a rescue therapy for pediatric ALI/ARDS in the last two decades.48 Exogenous surfactant seems a promising therapy in children with ALI/ARDS.49 The role of corticosteroids in the prevention and treatment of ARDS/ALI in children is not established. The area of the safe window within the pressure volume loop depends on the severity of lung injury.45 The more severe the ARDS the smaller the safe window. This implies a smaller difference between optimal PEEP and safe maximal PIP. It is the theoretical basis of the application of HFO ventilation in severe ARDS. Theoretically HFO ventilation fits more optimal in a small safe window compared to conventional lung protective ventilation, Figure 3.

Figure 3 Schematic depiction of theoretical basis of lung protective ventilation strategies. Loop A represents the position of a pressure volume loop, within the overall pressure volume curve, during mechanical ventilation before a recruitment maneuver. Airway pressure is increased to recruit non aerated lung regions (1). Then PEEP or mean airway pressure mpaw for HFO ventilation is lowered to the lowest possible level avoiding derecruitment, optimal PEEP/ mpaw (2). Loop B and C represent the position of the pressure volume loop within the safe window for conventional ventilation and HFO ventilation respectively. Minimizing tidal volume and/or maximizing peak inspiratory pressures avoids overdistension during inspiration. 

Page 16: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General introduction and outline

15

Several animal studies show that HFO ventilation reduces VILI.50,51 While early human studies identified improved oxygenation and safety associated with this mode.52-54 Currently there is limited evidence to evaluate HFO ventilation in terms of its effect on mortality and ventilator-free days. Several case series and a small number of randomized studies have demonstrated HFO ventilation to be safe and effective for improving oxygenation of patients with ARDS.53-60 These studies also demonstrate a potential for improved survival for a subgroup of patients transitioned early to HFO ventilation.53,61,62 Recently, sustained improvements in oxygenation have been found for patients with early-onset ARDS.63 Conceptually it can be argued that lung protective ventilation should aim at the lowest possible Vt. In HFO ventilation lowest Vt’s are reached at higher oscillatory frequencies.64 In ALI/ARDS compliance of the respiratory system is compromised. HFO ventilation can still provide adequate gas exchange at higher oscillatory frequencies in a non-compliant respiratory system.65

Lung protective ventilation and spontaneous breathing Lung protective ventilation and spontaneous breathing are combined in several partial ventilatory modes. Airway pressure release ventilation (APRV) and biphasic positive airway pressure (BIPAP) are modes that allow unrestricted spontaneous breathing during the ventilator cycle. In assist-control ventilation (A/C) only some breathing effort is necessary to trigger the ventilation and assist a breath. Both experimental models and clinical studies, emphasize the positive effect of spontaneous breathing during mechanical ventilation on the distribution of inflation and ventilation in the diseased lung. The most important factor responsible is the diaphragm. The diaphragm consists of an anterior tendon plate and a posterior muscle section. An active diaphragm lowers pleural pressure during inspiration. As a result transpulmonary pressure increases, enabling better aeration of dependent lung regions close to the diaphragm. In addition, during spontaneous breathing, the active dorsal muscle section of the diaphragm shifts the ventilation to the dorsal lung areas. A positive effect on lung aeration was already observed in several animal and clinical studies, when during mechanical ventilation spontaneous breathing contributed only 10–30% to the total minute ventilation.66-69 Spontaneous breathing during mechanical ventilation improves oxygenation, lowers need for sedatives, improves hemodynamics.66,67,70 There are two small single center trials that showed a benefit in terms of survival and ventilator free days.67,71

Page 17: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 1

16

Not all research is in favor of the preservation of spontaneous breathing during mechanical ventilation. Some studies suggest that use of neuromuscular blocking agents (NMBAs) in the early phase of ALI/ARDS may improve oxygenation and reduce inflammation. Gentle spontaneous breathing may be beneficial, it can be argued that vigorous spontaneous respirations are contraindicated for the injured lung. Forceful respiration efforts can impose stress on the lungs and aggravate VILI.72

Work of breathing

The main goals of mechanical ventilation are to restore gas exchange and reduce the work of breathing (WOB) by assisting respiratory muscle activity. The total work performed by a spontaneously breathing, intubated patient connected to a mechanical ventilator consists of physiologic and imposed components. When spontaneous breathing is maintained in a partial ventilatory mode it is essential to understand the different components of WOB.73 Physiologic work includes elastic work, work to overcome the elastic forces of the lungs and chest wall during inflation; and resistive work to overcome the resistance of the airways and pulmonary tissue to the flow of gas. The imposed WOB is the resistive work performed by a patient to breath trough the apparatus, i.e. the endotracheal tube, breathing circuit, and the work needed to trigger the ventilator. A reason for mechanical ventilation of a patient is unloading of the respiratory muscles. Imposed WOB however may equal or exceed physiologic work under some conditions.74 Measuring WOB is a useful approach to evaluate ventilator design and performance. WOB can be assessed by several methods. In general, the work performed during each respiratory cycle is mathematically calculated as the area on a pressure volume diagram: WOB = ΔPressure × ΔVolume, and is expressed in Joules per liter (J/l). This pressure volume loop is referred to as the Campbell diagram.75 By measuring the esophageal, tracheal, airway pressure and assessing the static chest wall compliance, the different components of WOB can be calculated in the Campbell diagram, Figure 4. In the case of ineffective respiratory efforts, that is, muscle contraction without volume displacement, WOB cannot be measured from the Campbell diagram, since this calculation is based on volume displacement. In this situation, measurement of the pressure time product (PTP) more accurately reflects the energy expenditure of the respiratory muscles.76,77 The PTP is the

Page 18: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General introduction and outline

17

product of the pressure developed by the respiratory muscles multiplied by the time of muscle contraction, expressed in cmH2O per second. Again by measuring the esophageal, tracheal, airway pressure and assessing the static chest wall compliance the different components of work can be assessed. Normal WOB for an a healthy adult is 0.3–0.6 J/l.73 An elevated WOB level during mechanical ventilation results in dyspnea and discomfort.78,79 The optimal workload for critically ill patients is unclear. Research focuses mainly on WOB in the weaning phase.80,81 A WOB level in the physiologic range, approximately 0.5 J/l in adults, seems to correspond with an optimal workload. Full unloading, reducing the WOB to zero, induces loss of respiratory muscles. Excessive respiratory muscle loading may cause muscle fatigue and weaning failure.82 The workload of 0.5 J/l seems not only optimal during ventilator weaning, but also in the acute phase of respiratory failure.83,84 There are only a small number of studies reporting normal WOB values for pediatric patients. Physiologic WOB in healthy children and adolescents (6−18 years) ranges between 0.1 and 0.6 J/l.85 In healthy preterm and full-term

Figure 4 Campbell diagram. Work of breathing (WOB) measured by the esophageal pressure. Resistive WOB, consisting of imposed and physiologic resistive WOB. Elastic WOB, work to overcome the elastic forces of the lungs and chest wall during inflation. WOB related to intrinsic PEEP. During inspiration energy is stored in the respiratory system enabling passive expiration. In for example severe airway obstruction expiration can become active, expiratory WOB. To partition different components of WOB static chest wall compliance is necessary. Chest wall compliance represents the pleural (esophageal) pressure obtained when muscles are totally relaxed and lung volume increases above functional residual capacity (FRC), measured under static conditions.

Page 19: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 1

18

infants the WOB range is 0.02–0.2 J/l.86 The optimal WOB during mechanical ventilation for pediatric patients is even more unclear.

Spontaneous breathing during high-frequency oscillatory ventilation

Preservation of spontaneous breathing during mechanical ventilation was not yet an issue during the development of the HFO ventilator (SensorMedics, 3100A/B, Yorba Linda, CA, USA) in the 1970s and 1980s. In 1980 Butler was one of the first to describe the results of experiments of ventilation by high-frequency oscillations in humans.87 The experiments were performed in four physician volunteers and twelve ventilated patients. Subjects were ventilated with high-frequency oscillations for one hour. The results demonstrated that high frequency small volume oscillations could achieve excellent gas during prolonged apnea. The group of Butler recognized the potential advantages of apneic HFO ventilation. HFO ventilation, without spontaneous breathing activity, caused very little repetitive distention of lung parenchyma and could in this way reduce lung injury. In the late 1970s and 1980s the suppressive effect of HFO ventilation on spontaneous breathing was well recognized. Different animal and human studies elucidated the mechanisms responsible.88-93 Suppression of respiratory drive is not simply a matter of CO2 elimination. By the Hering-Breuer inflation reflex, lung volume attributes to this suppression of spontaneous breathing. The Hering-Breuer reflex is a protective reflex and prevents over inflation of the lungs. Pulmonary stretch receptors (PSR) present in the smooth muscle of the airways respond to excessive stretching of the lung during large inspirations or high lung volumes. Once activated, the PSRs send signals to the centers that control breathing in the medulla oblongata. At increasing lung volumes inspiration is more inhibited and apnea may occur. During HFO ventilation apnea is seen when the set mpaw induces a high lung volume. In a clinical setting spontaneous breathing is normally seen during HFO ventilation.94 Unfortunately, with current HFO ventilator circuit design, vigorous respiratory efforts in large pediatric and adult patients may cause pressure swings that activate alarms, interrupt oscillations, and produce significant desaturations. Initial HFO ventilation trials in adults recommended muscular paralysis for this reason. The use of HFO ventilation for life sustaining gas exchange is counter-balanced by the need for heavy sedation and possible muscular paralysis.

Page 20: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General introduction and outline

19

Furthermore weaning from the HFO ventilator may be prolonged due to sedative and paralytic use.

Outline of this thesis

As outlined in this introduction, current understanding of the pathophysiology of acute respiratory distress syndrome, acute lung injury and ventilator induced lung injury has led to the development of different lung protective ventilation strategies. HFO ventilation is one of these strategies. Preservation of spontaneous breathing is for several reasons advocated in conventional lung protective ventilation strategies, as is explained in this chapter. Unfortunately spontaneous breathing is not well tolerated during HFO ventilation.

Aims of the studies The main objective of this thesis was to optimize the HFO ventilator design (SensorMedics, 3100A/B, Yorba Linda, CA, USA) to better tolerate spontaneous breathing.

When during mechanical ventilation spontaneous breathing is maintained, the mechanical ventilator and ventilator circuit impose workload to the patient. The bench test described in chapter 2 was performed to investigate which factors contribute to the imposed WOB using a SensorMedics 3100A/B HFO ventilator. A demand flow system was developed to overcome the imposed WOB caused by the HFO ventilator. The demand flow system had to respond to a subject’ spontaneous breathing efforts during high frequency oscillations. Chapter 3 provides a detailed description of the design and control of the demand flow system. In chapter 4 the demand flow system performance is evaluated in a bench test. Reliable testing of the demand flow system during HFO ventilation is only possible using a spontaneously breathing subject. Therefore demand flow system performance was evaluated in an animal model of moderate lung injury. The results of these animal experiments are summarized in chapter 5 and 6. Chapter 5 is focussed on work of breathing during HFO ventilation. The HFO ventilator was used in the standard configuration and with use of the demand flow system. In chapter 6 the effects of spontaneous breathing and muscular paralysis during HFO

Page 21: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 1

20

ventilation on regional lung characteristics were studied with the use of electrical impedance tomography (EIT).

Chapter 7 provides a review of published clinical experiences, animal and bench studies with HFO ventilation. It reflects on how the use of HFO ventilation might be improved in light of these experiences and our own opinion.

In chapter 8 the results are summarized and discussed, a simplified explanation of the demand flow system controller design is provided, conclusions are given and future perspectives are presented.

Page 22: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

Imposed work of breathing during high-frequency oscillatory ventilation: a bench study

Marc van Heerde Huib R. van Genderingen

Tom Leenhoven Karel Roubik

Frans B. Plötz Dick G. Markhorst

Critical Care 2006

Page 23: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 2

22

Abstract

Introduction: The ventilator and the endotracheal tube impose additional workload in mechanically ventilated patients breathing spontaneously. The total work of breathing (WOB) includes elastic and resistive work. In a bench test we assessed the imposed WOB using 3100 A/3100 B SensorMedics high-frequency oscillatory ventilators. Methods: A computer-controlled piston-driven test lung was used to simulate a spontaneously breathing patient. The test lung was connected to a high-frequency oscillatory (HFO) ventilator by an endotracheal tube. The inspiratory and expiratory airway flows and pressures at various places were sampled. The spontaneous breath rate and volume, tube size and ventilator settings were simulated as representative of the newborn to adult range. The fresh gas flow rate was set at a low and a high level. The imposed WOB was calculated using the Campbell diagram. Results: In the simulations for newborns (assumed body weight 3.5 kg) and infants (assumed body weight 10 kg) the imposed WOB (mean ± standard deviation) was 0.22 ± 0.07 and 0.87 ± 0.25 J/l, respectively. Comparison of the imposed WOB in low and high fresh gas flow rate measurements yielded values of 1.63 ± 0.32 and 0.96 ± 0.24 J/l (p = 0.01) in small children (assumed body weight 25 kg), of 1.81 ± 0.30 and 1.10 ± 0.27 J/l (p < 0.001) in large children (assumed body weight 40 kg), and of 1.95 ± 0.31 and 1.12 ± 0.34 J/l (p < 0.01) in adults (assumed body weight 70 kg). High peak inspiratory flow and low fresh gas flow rate significantly increased the imposed WOB. Mean airway pressure in the breathing circuit decreased dramatically during spontaneous breathing, most markedly at the low fresh gas flow rate. This led to ventilator shut-off when the inspiratory flow exceeded the fresh gas flow. Conclusion: Spontaneous breathing during HFO ventilation resulted in considerable imposed WOB in pediatric and adult simulations, explaining the discomfort seen in those patients breathing spontaneously during HFO ventilation. The level of imposed WOB was lower in the newborn and infant simulations, explaining why these patients tolerate spontaneous breathing during HFO ventilation well. A high fresh gas flow rate reduced the imposed WOB. These findings suggest the need for a demand flow system based on patient need allowing spontaneous breathing during HFO ventilation.

Page 24: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: a bench study

23

Introduction

Maintenance of spontaneous breathing in mechanically ventilated patients augments ventilation perfusion matching and cardiopulmonary function, reduces sedative requirement and shortens the intensive care stay.66,67,70,95-97 High-frequency oscillatory (HFO) ventilation is a useful ventilatory mode for neonatal application98,99 and it is gaining interest in both pediatric and adult intensive care.100-103 Neonatal and small pediatric patients can easily breathe spontaneously during HFO ventilation. Muscular paralysis is avoided and only mild sedation needs to be applied to tolerate ventilation and reduce stress. In larger children and adults, however, spontaneous breathing during HFO ventilation is usually not well tolerated because of patient discomfort. The sedation level often has to be high and even muscular paralysis may be necessary.94 We speculate that this discomfort is caused by a high imposed work of breathing (WOB). The imposed WOB is the work added to the physiologic WOB when patients breathe through a breathing apparatus. This includes work to overcome resistance added by the endotracheal tube, the breathing circuit and the humidification device, and work required to trigger the ventilator demand flow system. A physiologic WOB of 0.3–0.6 J/l is considered normal in a healthy adult.73 Depending on the ventilator settings, the imposed WOB can contribute as much as 80% to the total work of breathing.74 The imposed WOB is greatest during continuous positive airway pressure (CPAP), where the patient performs all the effort required to ventilate.104 HFO ventilation may in this respect be regarded as a super-CPAP system. In a physical sense, work is performed when a transmural pressure ptm changes the volume V of a distensible structure: W = ΔP x ΔV, most often expressed in Joules per liter (J/l). Applied to a breathing apparatus, the imposed WOB is calculated by integrating the pressure measured at the tracheal end of the endotracheal tube pett times the volume change: imposed WOB = ∫insp dpett dv. As inspiration is active and expiration is usually passive, only the inspiratory imposed WOB is generally considered. In a SensorMedics HFO ventilator (3100A or 3100B; SensorMedics, Yorba Linda, CA, USA), the imposed WOB is directly related to the breathing-related difference between the set mean airway pressure mpaw and the pett; the greater the difference, the greater the imposed WOB and thus patient effort. mpaw is regulated by a continuous fresh gas flow rate and an expiratory balloon valve. During inspiration of a patient, air is inhaled from the ventilator and the pett level drops. The magnitude of this drop is influenced by the fresh gas flow rate, the endotracheal tube size and the inspiratory flow rate.73

Page 25: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 2

24

In order to find a solution to better tolerate spontaneous breathing during HFO ventilation in large pediatric and adult patients, we performed a bench test, in which the inspiratory imposed WOB and pressure fluctuations in mpaw were assessed for newborn to adult simulations. We evaluated which factors contributed to the imposed WOB in the SensorMedics HFO ventilator: fresh gas flow rate, endotracheal tube size or inspiratory flow rate.

Materials and methods

Bench test set-up A custom-made artificial lung was used to simulate a spontaneously breathing subject with variable age, Figure 1a. This test lung consisted of a tube, 10 cm in diameter, with a computer-controlled piston. A sinusoid flow simulated inspiration

Figure 1 Schematic drawing of the experimental set-up. A, the ventilator circuit of a Sensor Medics 3100 A/B oscillator is connected to a piston-driven test lung by an endotracheal (et) tube. pett and paw, pressures in the test lung and in the ventilator circuit respectively. Flow is measured at the proximal end of the endotracheal tube. HFO ventilator, high-frequency oscillatory ventilator. B, simulated spontaneous breath. C, modified Campbell diagram, where a is the start of inspiration and b is the end of inspiration. The ‘inspiration’ area represents the imposed inspiratory work of breathing (WOB), calculated using the modified Campbell diagram. CDP, continuous distending pressure or set mpaw.

Page 26: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: a bench study

25

of spontaneous breathing, exponential decelerating flow expiration, Figure 1b. The test lung was connected to a HFO ventilator (3100A or 3100B; SensorMedics) with an endotracheal tube (Rüschelit, Rüsch, Kernen, Germany). Different patient circuits were used for each HFO ventilator (3100 A or 3100 B; SensorMedics). The same heated humidifier was used for both ventilators (MR225 humidification chamber; Fisher and Paykel, Auckland, New Zealand). The inspiratory airway flow and the expiratory airway flow in the endotracheal tube were measured with a hot-wire anemometer (Florian; Acutronic Medical Systems AG, Hirzel, Switzerland). The tidal volume Vt of spontaneous breathing was calculated by flow integration. The pett value was measured using the Florian respiration monitor. The pressure at the Y-piece paw in the ventilator circuit was measured using the unfiltered electronic signal of the internal pressure sensor of the HFO ventilator. Flow and airway pressures were sampled at 100 Hz and were stored on a laptop computer for off-line analysis.

Table 1 Spontaneous breathing simulation and ventilator settings

Newborn Infant Small child Large child Adult

Assumed weight (kg) 3.5 10 25 40 70

Spontaneous breathing simulation:

Respiratory rate (/min) 35 30 25 20 12 Vt (ml/kg) / peak flow (l/min) 5 / 2.6 5 / 6.3 5 / 13 5 / 17 5 / 18

7 / 3.6 7 / 8.9 7 / 19 7 / 24 7 / 25 10 / 5.1 10 / 12 10 / 27 10 / 34 10 / 36

Inspiratory/expiratory ratio 1 : 2 1 : 2 1 : 2 1 : 2 1 : 2

Ventilator settings:

HFO ventilator A A B B B

Tube size (mm) 3.0, 3.5, 4.0 4.0, 4.5, 5.0 5.5, 6.0, 6.5 6.5, 7.0, 7.5 7.5, 8.0, 8.5

Fresh gas flow rate (l/min) 15 / 20 20 / 40 20 / 60 20 / 60 20 / 60

CDP (cmH2O) 18 25 25 25 25

ΔP (cmH2O) 35 50 50 50 50

Oscillation frequency (Hz) 10 8 8 6 6

Vt, tidal volume; peak flow, peak inspiratory flow; HFO ventilator, SensorMedics 3100; CDP, continuous distending pressure; ΔP, proximal pressure amplitude.

Page 27: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 2

26

The HFO ventilator was set to a specific patient size as prescribed by the operator's manuals for management of acute respiratory distress syndrome (ARDS).105,106 We tested five patient weight ranges, from newborn to adult, Table 1. The ventilator fresh gas flow rate was set at two different levels: low and high. For all different patient sizes, three Vt levels of normal spontaneous breathing were simulated. The peak inspiratory flow rates that were generated with these Vt levels are also presented in Table 1. Three different sizes of endotracheal tubes were used for each patient size, Table 1. In total, 90 different settings were tested.

Imposed work of breathing For each experimental condition, 12–20 breaths were recorded. The inspiratory imposed WOB was calculated for each simulated spontaneous breath, based on the modified Campbell diagram, Figure 1c:107

Imposed WOB = ∫insp (CDP – mpett) dv where CDP is the continuous distending pressure or set mpaw level on the SensorMedics oscillator, and mpett is the mean airway pressure in the test lung. This was calculated by low-pass filtering (Butterworth filter with a cutoff frequency of 10 Hz) of the pett signal to eliminate pressure changes on account of oscillations. The imposed WOB was averaged over all breaths (expressed as J/l).

Airway pressure Swings of the pressure in the ventilator circuit due to oscillations were removed by low-pass filtering. As a result, all changes in airway pressure were attributable to the settings chosen to mimic spontaneous ventilation. Pressure fluctuations due to spontaneous breathing Δmpaw are expressed as the deviation from set CDP (in cmH2O). Δmpaw,insp is the maximum deviation from the set CDP during inspiration, and Δmpaw,exp is the maximum deviation from the set CDP during expiration. Δmpaw,insp and Δmpaw,exp were calculated separately as the inspiratory and expiratory flow patterns of spontaneous breathing differed.

Statistical analysis Data are expressed as the mean ± standard deviation. Comparison of means for normally distributed data was performed with an independent t-test. p < 0.05 was considered statistically significant. Linear regression was performed to explore relations between the imposed WOB, the endotracheal tube size, the fresh gas flow

Page 28: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: a bench study

27

rate and the peak inspiratory flow. Statistical analyses were performed using SPSS 11.5 for Windows (SPSS Inc., Chicago, IL, USA).

Results

Imposed work of breathing The imposed WOB was 0.22 ± 0.07 J/l for all measurements in the newborn (assumed body weight 3.5 kg) simulations and was 0.87 ± 0.25 J/l in the infant (assumed body weight 10 kg) simulations, Figure 2. Linear regression showed that a high or a low fresh gas flow rate did not independently influence the imposed WOB in these measurements (p = 0.64 for newborns and p = 0.94 for infants) (3100 A). An independent contributor to the imposed WOB was the peak inspiratory flow; a higher peak inspiratory flow increased the imposed WOB (p < 0.001). The tube size did not independently contribute to the imposed WOB (p = 0.92 for newborns and p = 0.92 for infants). The imposed WOB for the larger pediatric and adult patient size simulations (3100 B oscillator; SensorMedics) was significantly higher in the low fresh gas flow rate condition in comparison with the high fresh gas flow rate condition. The results for the imposed WOB for low flow versus high flow were 1.63 ± 0.32 versus 0.96 ± 0.24 J/l (p = 0.01) in the small child (assumed body weight 25 kg) simulation,

Figure 2 Imposed work of breathing. Imposed work of breathing (WOBi) for all simulations. Results for all measurements with the 3100 A and 3100 B SensorMedics oscillators.

Page 29: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 2

28

were 1.81 ± 0.30 versus 1.10 ± 0.27 J/l (p < 0.001) in the large child (assumed body weight 40 kg) simulation, and were 1.95 ± 0.31 versus 1.12 ± 0.34 J/l (p < 0.001) in the adult (assumed body weight 70 kg) simulation. Independent contributors to the imposed WOB were the fresh gas flow rate (p < 0.001) and the peak inspiratory flow (p < 0.001). A high fresh gas flow rate decreased the imposed WOB, and a high peak inspiratory flow increased the imposed WOB. The tube size did not independently contribute to the imposed WOB (p = 0.07).

Airway pressure The mpaw in the ventilator circuit decreased dramatically during spontaneous breathing, most markedly at a low fresh gas flow rate, Figure 3. In this example the mpaw in the ventilator circuit even becomes negative. This effect was observed when the fresh gas flow rate was low and with a Vt of 7 or 10 ml/kg for the large child and adult patient simulations. In these simulations the peak inspiratory flow exceeded the fresh gas flow rate. This triggered the automatic ventilator shut-off, a safety feature of the SensorMedics oscillator.

Table 2 Maximum deviation of mean airway pressure from the set continuous distending pressure

Δmpaw,insp (cmH2O) Δmpaw,exp (cmH2O)

Low flow a High flow b p value Low flow a High flow b p value

SensorMedics 3100A

Newborn 2.17 ± 0.24 1.75 ± 1.04 NS 0.83 ± 0.27 0.71 ± 0.44 NS

Infant 6.85 ± 1.34 4.56 ± 1.24 0.002 3.73 ± 1.77 3.59 ± 1.7 N�

SensorMedics 3100B

Small child 17.0 ± 3.11 8.91 ± 2.11 < 0.001 8.90 ± 2.49 5.21 ± 1.70 0.002

Large child 23.0 ± 2.60 12.8 ± 2.62 < 0.001 13.2 ± 2.01 8.76 ± 1.66 < 0.001

Adult 25.5 ± 2.51 13.7 ± 3.50 < 0.001 15.9 ± 2.97 8.34 ± 2.18 < 0.001

Δmpaw,insp , maximum deviation of mean airway pressure from continuous distending pressure during inspiration; Δmpaw,exp , maximum deviation of mean airway pressure from continuous distending pressure during expiration. aLow fresh gas flow rate: 3100A ventilator, 15 l/min and 3100B ventilator 20 l/min. bHigh fresh gas flow rate: 3100A ventilator, 20 l/min and 3100B ventilator 60 l/min

Page 30: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: a bench study

29

Δmpaw,insp and Δmpaw,exp for all measurements in the newborn (assumed body weight 3.5 kg) simulations were not significantly different comparing low and high fresh gas flow rates, Table 2. In the infant (assumed body weight 10 kg) simulations the Δmpaw,insp value was significantly lower in the high fresh gas flow rate condition in comparison with the low fresh gas flow rate testing (p = 0.002). There was no difference in Δmpaw,exp measurements (3100 A; SensorMedics). For pediatric and adult simulations (3100 B) the Δmpaw,insp and Δmpaw,exp values were significantly lower in the high fresh gas flow rate condition in comparison with the low fresh gas flow rate condition, Table 2.

Discussion

The main result of this study is that the imposed WOB can be markedly increased during HFO ventilation in pediatric and adult patient simulations, especially at low fresh gas flow rates. This can be a good explanation for the discomfort seen in patients breathing spontaneously during HFO ventilation. The fresh gas flow rate

Figure 3 Example of the changes in pressures caused by spontaneous breathing. Example of the changes in paw and mpaw during the simulation for a large child (assumed body weight 40 kg, tidal volume 280 ml), for both low and high fresh gas flow rates. Note that pressure changes decrease with a higher fresh gas flow rate and thus the imposed work of breathing decreases. In this example at the low fresh gas flow rate, the pressure in the ventilator circuit becomes negative as the inspiratory flow exceeds the fresh gas flow rate (arrow). paw, unfiltered airway pressure; mpaw , filtered airway pressure calculated by low-pass filtering of the paw signal; CDP, continuous distending pressure or set mpaw; flow tube, filtered flow over the endotracheal tube.

Page 31: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 2

30

and peak inspiratory flow are both strongly related to the imposed WOB. The mpaw is not maintained in the breathing circuit when inspiratory flow exceeds the fresh gas flow rate, and this can even lead to ventilator shutdown.

Work of breathing Compared with the WOB of a healthy adult (0.3–0.6 J/l), the imposed WOB is high if spontaneous breathing is simulated during HFO ventilation.73 As the physiologic WOB is not considered in this bench test, the total WOB is even higher in a patient breathing spontaneous during HFO ventilation. An elevated WOB level results in dyspnea and discomfort.78,79 The optimal workload for critically ill patients is unclear. Research focuses mainly on WOB in the weaning phase.80,81 A WOB level in the physiologic range (approximately 0.5 J/l in adults) seems to correspond with an optimal workload. Full unloading (for instance, reducing the WOB to zero) induces loss of respiratory muscles. Excessive respiratory muscle loading may cause muscle fatigue and weaning failure.82 The workload of 0.5 J/l seems not only optimal during weaning, but also in the acute phase of respiratory failure.83,84 In the pediatric and adult simulations, the imposed WOB exceeded the normal physiologic WOB by as much as 400%. There are very few studies reporting normal WOB values for pediatric patients. WOB in healthy children and adolescents (6−18 years) ranges between 0.1 and 0.6 J/l.85 In healthy preterm and full-term infants the WOB range is 0.02–0.2 J/l.86 The optimal WOB during mechanical ventilation for these patients is even more unclear. Our results show that the level of imposed WOB is high during spontaneous breathing in HFO ventilation. A high fresh gas flow rate in simulations for pediatric and adult patients reduces the imposed WOB, but not within the physiologic range of WOB. It seems logical to aim at a level of imposed WOB in the physiologic WOB range. However, there are no data to support this. An effective way to reduce imposed WOB in HFO ventilation is desirable. Although we did not simulate this condition, a possible solution is to set the fresh gas flow rate to a higher rate. Another solution is the use of a demand flow system instead of the continuous fresh gas flow rate. In order to reduce the imposed WOB, the fresh gas flow rate has to far exceed the peak inspiratory flow. A reasonable suggestion would be the possibility to generate peak fresh gas flow rates comparable with conventional ventilation (approximately 140 l/min) depending on patient need. Since the imposed WOB does not reflect the isovolumetric breathing effort, the pressure time product per breath was also calculated (data not included). As expected, results for the pressure time product per breath and the imposed WOB

Page 32: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: a bench study

31

were identical. This is explained by the lung model we used for spontaneous breathing, in which the inspiratory and expiratory flows were programmed. The volume changes were imposed, so isovolumetric contraction did not occur. For the simulations for newborns and infants, the imposed WOB was not influenced by the fresh gas flow rate. This may be explained by the chosen small difference in levels of low and high fresh gas flow rate. Simulations for newborns show a low level of imposed WOB. This is in agreement with the fact that these patients tolerate spontaneous breathing during HFO ventilation. Various factors define the imposed WOB. The endotracheal tube, the breathing circuit, the humidification device and the trigger settings impose the workload. Endotracheal tubes have the greatest effect on flow-resistive work.80 This seems in contrast with our results, and is explained by the small differences in tube sizes used in our experiments, relative to the large variations in peak inspiratory flow.

Airway pressure Large fluctuations in the mpaw in the breathing circuit are responsible for a high imposed WOB. They may also lead to unwanted alarms of the ventilator during HFO ventilation, or even to shut down. Upper and lower alarm limits are routinely set 3–5 cmH2O above and below the desired mpaw.105,106 This is a safety precaution against unnoticed mpaw changes due to changes in respiratory system compliance, which may lead to alveolar derecruitment or overdistension. Airway pressure fluctuations exceeded the alarm limit of 5 cmH2O in all simulations on the SensorMedics 3100 B ventilator. In the simulations for smaller patient size, alarm limits of 3 cmH2O above and below the CDP were sufficient to avoid alarms – although in most other measurements the alarm limits had to be set wide to stop alarms, interfering with patient safety in a clinical setting. If the peak inspiratory flow exceeded the fresh gas flow rate, this led to ventilator shutdown. This triggered the automatic ventilator shut-off, a safety feature of the SensorMedics oscillator.

Limitations of the study The imposed WOB is strongly related to the choice of Vt, respiratory rate, breathing pattern and tube size. In this in vitro study we aimed to choose realistic test conditions. However, in vivo conditions may differ from our bench test. In the lung model used only the imposed WOB can be evaluated. Patient or total work cannot be assessed. During HFO ventilation the lungs can expand to high levels of end expiratory lung volume near the total lung capacity. At high levels of end

Page 33: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 2

32

expiratory lung volume the work of breathing will increase as a result of an increase in elastic work. The Vt levels used for simulations were fixed. The Vt level that a patient can generate is influenced by the level of end expiratory lung volume. Only shallow breathing is possible at levels of end expiratory lung volume near the total lung capacity. In this lung model we are not able to evaluate these effects on patient WOB and on total WOB. These findings need validation in clinical practice.

Conclusions

The imposed WOB is considerable in spontaneous breathing in pediatric and adult patients during HFO ventilation and is a good explanation for the observed patient discomfort. A high fresh gas flow rate decreased the imposed WOB, but not sufficiently. Large swings in airway pressure complicate the setting of safe alarm limits and can even lead to ventilator malfunction. In clinical practice it is reasonable to consider the level of the fresh gas flow rate. In order to minimize the imposed WOB and to allow at least shallow spontaneous breathing during HFO ventilation, the fresh gas flow rate has to be set at a high level.

Page 34: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

Design and control of a demand flow system for a high-frequency oscillatory ventilator

Karel Roubik Jakub Ráfl

Marc van Heerde Dick G. Markhorst

IEEE Transactions on Biomedical Engineering 2011

Page 35: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

34

Abstract

Lung protective ventilation is intended to minimize the risk of ventilator induced lung injury (VILI) and currently aimed at preservation of spontaneous breathing during mechanical ventilation. High-frequency oscillatory (HFO) ventilation is a lung protective ventilation strategy. Commonly used HFO ventilators, SensorMedics 3100, were not designed to tolerate spontaneous breathing. Respiratory efforts in large pediatric and adult patients impose a high workload to the patient and may cause pressure swings that impede ventilator function. A demand flow system (DFS) was designed to facilitate spontaneous breathing during HFO ventilation. Using a linear quadratic Gaussian state feedback controller, the DFS alters the inflow of gas into the ventilator circuit, so that it instantaneously compensates for the changes in mean airway pressure mpaw in the ventilator circuit caused by spontaneous breathing. The undesired swings in mpaw are thus eliminated. The DFS significantly reduces the imposed work of breathing and improves ventilator function. In a bench test the performance of the DFS was evaluated using a simulator ASL 5000 testlung. With the gas inflow controlled, trigger delay was 50 ms and mpaw was returned to its preset value within 115 ms after the beginning of inspiration. The trigger performance of the DFS is similar to the performance of modern conventional ventilators. The DFS might help to spread the use of HFO ventilation in clinical practice.

Page 36: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

35

Introduction

Mechanical ventilation (MV) is an effective, life-saving technique for the management of patients with respiratory failure. Nevertheless, numerous studies have shown that MV itself can initiate as well as exacerbate lung injury. This is known as ventilator induced lung injury (VILI).6 It led to the development of lung protective MV strategies that can protect the already injured lung from additional harm. An important issue is not to fully take over the patients’ respiration, but to allow spontaneous breathing. Spontaneous breathing during MV has clinical relevant positive effects on the diseased lung, lowers need for sedatives, and shortens duration of both MV and intensive care stay.108 The current trend in MV is to synchronize the ventilator’s activity with the breathing of a patient. During the last decades, advances in control technology have played a major role in the evolution of MV.109 Much effort was made to adapt ventilators to a spontaneously breathing patient. The incorporation of advanced control in modern ventilators has moved the control of ventilation from the machine to the patient. New MV modes provide automatic control of ventilator output in response to patients’ changing requirements. Algorithms for the assessment of successful weaning are being sought,110 and some ventilators even use optimal control algorithms to automatically wean patients from MV.109 Research activities are also aimed at development of closed-loop MV controllers with explicit definition of target physiological parameters, such as blood oxygen saturation and end-tidal concentration of CO2.111 High-frequency oscillatory (HFO) ventilation is a lung protective strategy very different from other modes of MV. Whereas the physiologic breathing frequency is 0.2–0.35 Hz (12–21 breaths per minute) during quiet breathing, HFO ventilation delivers pressure oscillations of 3–15 Hz around a constant mean airway pressure mpaw, producing tidal volumes Vt of approximately 1–2 ml/kg body weight. These Vt’s are often less than anatomical dead space, and associated swings in alveolar pressure are very small.64 This approach should theoretically limit VILI.42 Unfortunately, ventilators for HFO ventilation still do not offer such sophisticated control of MV as other ventilators. Spontaneous breathing of a patient is difficult during HFO ventilation, which represents one of the most significant disadvantages of HFO ventilation. The SensorMedics 3100 HFO ventilator (SensorMedics, Yorba Linda, CA, USA) was developed in the early 1980s and is the most commonly used HFO ventilator in large pediatric and adult patients. These HFO ventilators use open-loop control for

Page 37: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

36

all ventilator settings. Preservation of spontaneous breathing during MV was not taken into account during the design and causes two main problems. First, vigorous respiratory efforts in large pediatric and adult patients impedes HFO ventilator function. It causes pressure swings that activate alarms, interrupt oscillations and produce significant oxygen desaturations. Initial HFO ventilation trials in adults recommended muscular paralysis for this reason.94 Second, when a patient breathes spontaneously during MV, all components of the ventilator circuit impose a resistive workload to the patient. This imposed workload caused by the HFO ventilator is rather high and makes it difficult to breathe during HFO ventilation.112 Nonetheless, current protocols attempt to maintain spontaneous breathing during HFO ventilation.113 The aim of this study is to eliminate the disadvantage of HFO ventilation, the difficulties of spontaneous breathing during HFO ventilation. This chapter presents the design, control, and function of a device designed to facilitate spontaneous breathing during HFO ventilation with SensorMedics 3100 HFO ventilators. The chapter is organized as follows: First the problem of spontaneous breathing during HFO ventilation with the standard HFO ventilator is specified and the proposed solution is introduced. Then the new configuration of the HFO ventilator is described and the control algorithm is explained. The system performance is evaluated in a bench test. Finally, discussion on the system performance, limitations and the study conclusions are given.

Problem formulation

A standard configuration of a SensorMedics 3100 HFO ventilator is presented in Figure 1a. The inspiratory gas mixture from an air-oxygen blender enters a bias flow generator in the HFO ventilator. The manually adjustable generator can deliver a fixed bias flow rate Qbf between 0 and 60 l/min. This flow passes through the ventilator circuit and escapes via a balloon valve. The balloon valve represents a resistance, which, together with the bias flow, determines mpaw in the ventilator circuit. The mpaw assures a sufficient alveolar lung surface area for gas diffusion and controls oxygenation. The generation of inspiratory and expiratory volume is provided by an oscillating membrane, operating at a frequency between 3 and 15 Hz. The oscillating membrane generates pressure swings that propagate through the ventilator circuit to the airway opening of the patients’ respiratory

Page 38: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

37

system. These changes of pressure create a bidirectional flow of gas between the ventilator circuit and the connected airways. The relatively low and fixed Qbf is the reason that spontaneous breathing is not well tolerated during HFO ventilation.112 The fixed Qbf results in a decrease in mpaw during spontaneous inspiration and an increase in mpaw during expiration. A ventilator may evaluate the pressure swings caused by the patients’ breathing as a possible danger. Patients’ exhalation induces an increase in mpaw in the ventilator circuit which may exceed the set alarm limits. As a result, the alarm is activated. The same may occur during a patients’ inhalation when a decrease in mpaw in the ventilator circuit may be considered as a significant gas leak or circuit disconnection. This can even lead to automatic ventilator shut off, a safety feature of the HFO ventilator. A patients’ spontaneous breathing that does not affect ventilator performance may be assured by an additional device which instantaneously compensates for the swings in mpaw induced by spontaneous breathing. Such a device should rapidly react to the changes in the amount of gas in the ventilator circuit. When using a fixed Qbf, this amount of gas decreases as a result of a patients’ inhalation and increases as a result of a patients’ exhalation. A device that is able to compensate for

Figure 1 High-frequency oscillatory ventilator setups. The high-frequency oscillatory (HFO) ventilator connected to a patient or test lung in the standard configuration with constant bias flow (a) and with the demand flow system providing the variable gas inflow (b). paw, airway pressure; Qbf, fixed bias flow rate; qDFS, bias flow rate; DFS, demand flow system.

Page 39: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

38

these volume changes in the ventilator circuit would promptly eliminate the undesired pressure swings in mpaw. We refer to such a device as the demand flow system (DFS). There are two main advantages of the DFS. First, full compensation of the gas volume changes in the ventilator circuit completely eliminates mpaw changes in the ventilator circuit caused by a patients’ spontaneous breathing. As mpaw is the only parameter monitored by the HFO ventilator, for both operational and safety reasons, functioning of the HFO ventilator will not be affected by spontaneous breathing activity. Next, rapid compensation of gas volume changes would significantly reduce a patients’ breathing workload. When spontaneous breathing of a patient connected to a HFO ventilator is desired, the imposed work of breathing (WOB) must be reduced. The DFS may reduce the imposed workload significantly, as it easily makes available any amount of inspiratory gas upon a patients’ demand. Hence, the patient needs less effort to overcome the high resistance components of the original HFO ventilator circuit.

Demand flow system design

The DFS replaces the original fixed Qbf through the ventilator circuit by gas flow instantaneously adjusted according to a spontaneously breathing patient. In the new configuration, Figure 1b, the HFO ventilator provides only the oscillations, assuring bidirectional gas flow between the ventilator circuit and a patiens’ lung, whereas gas flow through the ventilator circuit is delivered by the

Figure 2 The fundamental structure of the demand flow system. paw, airway pressure; qDFS, bias flow rate; A/D, analog-to-digital converter; D/A digital-to-analog converter.

Page 40: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

39

DFS. This system maintains the necessary bias flow and simultaneously eliminates mpaw changes in the ventilator circuit caused by spontaneous breathing. When a patient starts to inhale, gas is removed from the ventilator circuit, which leads to a drop in mpaw, detected by a pressure sensor. In response, the DFS increases the gas flow rate qDFS into the ventilator circuit so that the flow through the expiratory balloon valve is returned to its original value, and therefore mpaw in the circuit is maintained constant. Similarly, reduction of gas inflow into the circuit maintains mpaw unaltered during spontaneous exhalation. As a result, mpaw does not change regardless of a patients’ spontaneous breathing effort. The fundamental structure of the DFS is depicted in Figure 2. The pressure sensor (14PC03D, Honeywell, USA) converts the proximal airway pressure paw, measured at the proximal end of the endotracheal tube, to an analog voltage signal. The voltage signal is sampled and digitized by an A/D converter in a data acquisition board (DAQCard-6024E, National Instruments Corporation, Austin, TX, USA) and sent to a control computer, where it is converted into a discrete proximal airway pressure signal paw (k) at a discrete time k. A control algorithm (described in the next section) realized in the MATLAB/Simulink environment (The MathWorks, Natick, MA, USA) calculates the desired discrete value of the gas flow rate qDFS (k). The voltage necessary for the proportional valve control is derived from the qDFS (k) value using the conversion curve of the proportional valve. The discrete value of the control voltage is converted into an analog voltage control signal by the same data acquisition board. After its D/A conversion, the analog voltage signal enters a micro driver that actuates the proportional valve so that the desired gas flow rate qDFS can be delivered into the ventilator circuit. The valve, with a transmission time of 30 ms (10–90%), was obtained from an AVEA mechanical ventilator (CareFusion, Yorba Linda, CA, USA). The DFS operates in two modes referred to as MANUAL and DEMAND. In the MANUAL mode, the mpaw value is acquired as a 3 s moving average of paw.. This mpaw value serves as the target mpaw value for the regulator during the DEMAND mode of operation. In the MANUAL mode the fresh gas flow rate is constant; it is set by a potentiometer. This value of the flow rate is stored as the static Qbf when the DFS is switched to the DEMAND mode. In the DEMAND mode, the variable component of the gas flow is automatically computed based on the difference between the current paw and the target mpaw value. The proportional valve is adjusted so that the total flow rate into the system is the sum of the static Qbf rate and the calculated variable flow rate component.

Page 41: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

40

Demand flow system control

The DFS controls the pressure in the ventilator circuit by means of the variable gas inflow into the circuit. The control program calculates the discrete value of the gas flow rate qDFS(k), based on the knowledge of the discrete proximal airway pressure signal paw(k). A discrete-time linear quadratic Gaussian (LQG) state feedback controller was employed; i.e., a combination of a linear quadratic (LQ) state-feedback regulator and a state observer estimating the unknown state of a system with Gaussian stochastic disturbances. The controller is designed as a regulator that rejects the disturbances to the paw(k) signal caused by spontaneous breathing. The selected state-space approach for control requires an internal model of the regulated system, of which the main part is a model of a patient-ventilator system (the process model). The state-variable representation of the process was derived for the time-varying components of the gas flow rate qDFS and the proximal airway pressure paw. A discrete-time version of this arrangement was utilized by the LQG controller.

Process Model A simple linear lumped-parameter model of the regulated patient-ventilator system is presented in Figure 3. The model was created using an electro-acoustic analogy. In this model, a resistor–capacitor (RC) network, together with pressure and gas-flow sources, is used to approximate the behavior of a patients’ respiratory system, the HFO ventilator, and the ventilator circuit.114 A patients’ respiratory system is modeled by resistor R1, representing the sum of airway and endotracheal tube resistance, and by capacitor C1, representing the respiratory system compliance.

Figure 3 Model of the regulated patient-ventilator system. See text for explanation.

Page 42: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

41

Spontaneous breathing is modeled by the source of pressure pspont. Mechanical properties of a HFO ventilator are described by resistor R2, representing the airflow resistance of the ventilator chamber, and by capacitor C2, representing its compliance. The oscillations generated by the ventilator are introduced into the model using the pressure source pHFO. Concerning the ventilator circuit, the most significant mechanical property is exhibited by the expiratory balloon valve, which is, together with the expiratory circuit resistance, modeled by resistor R3. Gas inflow into the ventilator circuit from the DFS is modeled by the controlled gas flow source qDFS. The pressure nodes between their respective resistors and capacitors are denoted as p1 and p2. Finally, the last remaining node in the model refers to the proximal airway pressure paw. For the pressures p1, p2, and paw, the node equations can be derived in form:

0d

)d( spont11

1

aw1 =−

+−

tpp

CR

pp (1)

0d

)d( HFO22

2

aw2 =−

+−

tppC

Rpp

(2)

03

aw

2

2aw

1

1awDFS =+

−+

−+−

Rp

Rpp

Rppq . (3)

As has already been mentioned, the constant gas flow qDFS with the output resistor R3 assure the constant value of mpaw when a stable state is not disturbed by a patients’ spontaneous respiratory effort. That is, paw = mpaw = R3Qbf when qDFS = Qbf, where Qbf is the static bias flow. However, in a general case, the actual value of the proximal airway pressure paw may differ from mpaw by Δpaw:

awbf3awawaw pQRpmpp Δ+=Δ+= . (4)

Accordingly, p1 and p2 may be expressed as:

1bf31aw1 pQRpmpp Δ+=Δ+=

2bf32aw2 pQRpmpp Δ+=Δ+= . (5)

The actual value of the gas flow rate qDFS controlled by the DFS may also differ from the static bias flow:

Page 43: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

42

DFSbfDFS qQq Δ+= . (6)

If (4)–(6) are substituted into (1)–(3), the system equations for a state vector Δp12 = [Δp1 Δp2]T can be derived as: wGpFp +Δ+Δ=Δ DFS1212 q& (7)

DFS12 qRp aw Δ+Δ=Δ pH . (8)

where

⎥⎥⎥⎥

⎢⎢⎢⎢

=

2

222

2

21

1

21

1

121

GCGGG

GCGG

GCGG

GCGGG

F ,

T

2

2

1

1⎥⎦

⎤⎢⎣

⎡=

GCG

GCGG , T

HFOspont ][ pp &&=w , ⎥⎦⎤

⎢⎣⎡=

GG

GG 21H ,

GR 1

= , 1

11R

G = , 2

21

RG = ,

33

1R

G = , 321 GGGG ++= .

The actual values of parameters used in this model are presented in the appendix. Equations (7), (8) describe a linear, time-invariant, continuous-time system with the input variable flow ΔqDFS and the output pressure Δpaw, which is the difference between the mpaw and the actual paw. The bias flow Qbf does not influence the state vector pΔ12. The system state is, however, affected by the first time derivatives of a patients’ spontaneous breathing and high-frequency oscillations, i.e., by the vector w. The controller handles w as process noise causing stochastic perturbations in the system state. The continuous-time state-space model given by (7) and (8) was discretized with a sampling rate of TS = 200 Hz. The equivalent discrete-time equations are:

)()()()1( DFS1212 kkqkk wΓpΦp +Δ+Δ=+Δ (9)

)()()( DFS12aw kqRkkp Δ+Δ=Δ pH (10)

where

T2112 ][ )()()( kpkpkp ΔΔ=Δ , { }sexp TFΦ = , and { } GFΓ

T⎟⎠⎞

⎜⎝⎛= ∫ ττ dexps

0

Page 44: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

43

The matrices Φ and Γ were derived from matrices F and G in a standard way by solving (7) for discrete time points kTs and (k + 1)Ts, and further assuming that ΔqDFS is constant between the time points.115

Controller design The paw signal consists of three principal components; two of them are variable in time and one is constant. The variable pressure components are the high-frequency oscillations produced by the HFO ventilator and the pressure swings caused by spontaneous breathing. These patient-induced pressure swings contain predominantly lower frequency components than the high-frequency oscillations. The two variable pressure components are superimposed on a steady value of mpaw. The control algorithm separates the constant mpaw value and the high-frequency oscillations from paw(k). The remaining pressure signal comprises the spontaneous breathing component only and has to be suppressed by the controller. The implemented signal-processing procedure can be followed in Figure 4. First, Δpaw(k) is computed by subtracting the known target value of mpaw from the discrete-time output of the process paw(k). Then, the high-frequency pressure oscillations are removed using a low-pass, second order discrete Butterworth filter with a cutoff frequency of 1.5 Hz. The discrete-time output signal e(k) of the filter represents the undesired perturbation in the proximal airway pressure signal due to the spontaneous breathing. The objective of the LQ regulator with the feedback gain vector of −K is to minimize the filter output signal e(k).

Figure 4 Block diagram of the demand flow system control algorithm. See text for explanation.

Page 45: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

44

The LQ regulator calculates the variable discrete flow-rate component ΔqDFS(k) that minimizes the quadratic loss function JLQ:

( )∑∞

=

ΔΔ+=0

DFSqT

DFSeT

LQ )()()()(k

kqQkqkeQkeJ (11)

where Qe and Qq are positive weighting parameters. A one-step delay block was placed after the LQ gain block to prevent an algebraic loop. The desired discrete value of the total gas flow rate is calculated as qDFS(k) = Qbf + ΔqDFS(k – 1). The system controlled by the LQG controller is a serial connection of the delay block, the patient-ventilator system, and the low-pass filter. In the process model, mpaw, determined by Qbf and R3, is completely separated from proximal airway pressure swings Δpaw. Therefore, the steady components mpaw and Qbf are not relevant for the process of regulation, and for the LQG controller the patient-ventilator system is represented by (9) and (10). The LQ regulator requires the full state information of the controlled system. As the process variables p1 and p2 cannot be measured and as the state Δp12(k) is subject to the stochastic perturbation w(k) according to (9), a linear state estimator, the Kalman filter, is used to estimate Δp12(k) from the process input and the process output as Δ p̂ 12(k | k). The separation theorem guarantees that the LQ regulator can be designed independent of the Kalman filter as if the full-state information was available.115 For the LQ regulator, the process is represented by the deterministic part of (9), i.e. without the vector w(k). The system controlled by the LQ regulator can be described by the following: )()()1( DFS kqkk Δ+=+ BAxx (12)

)()( kke Cx= (13)

where

⎥⎥⎥

⎢⎢⎢

=

LPFLPFLPF

000

AHBB0ΦΓA

R

, T]00001[=B ,

TLPFLPFLPF ][ CHDDC R= , 0 is null matrix.

The matrices ALPF, BLPF, CLPF and DLPF originate in the state-space representation of the low-pass Butterworth filter. The state vector x(k) consists of the state of the

Page 46: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

45

delay block xd(k) = ΔqDFS(k – 1), the process state estimate Δ p̂ 12(k|k), and the low-pass filter state xLPF(k): x(k) = [ΔqDFS(k – 1) Δ p̂ 12(k|k) xLPF(k)]T. The loss function (11) can thus be rewritten as:

( )∑

=

ΔΔ+=0

DFSqT

DFST

LQ )()()()(k

kqQkqkkJ Qxx . (14)

where Q = CTQeC is a positive symmetric semidefinite weighting matrix. A solution to the LQ problem is a linear control law in the form of:

)()(DFS kkq xK−=Δ (15)

The regulator constant gain −K is calculated from:

SABSBBK T1Tq ][ −+= Q (16)

where S is a solution of the algebraic Riccati equation:115,116

QSABSBBSBASAAS ++−= − T1Tq

TT ][Q . (17)

Values of the weights used for the LQ regulator are presented in the appendix. The Kalman filter generates the a posteriori estimate Δ p̂ 12(k|k), which in each time step minimizes the criterion:

( ) ( )( )( ){ }

{ })|(Tr)|(ˆ)()|(ˆ)(ETr

)|(ˆ)()|(ˆ)(E

][][

T12121212

1212T

1212KF

kkkkkkkk

kkkkkkJ

Ppppp

pppp

=

Δ−ΔΔ−Δ=

Δ−ΔΔ−Δ=

(18)

where Tr{} denotes the matrix trace operator and P(k|k) is the covariance of the estimation error of Δ p̂ 12(k|k). The Kalman filter assumes a stochastic linear discrete-time single-input single-output system given by (9) and by:

)()()()( DFS12aw kvkqRkkp +Δ+Δ=Δ pH , (19)

with w(k) in (9) and v(k) in (19) being white, zero-mean, and mutually independent discrete-time Gaussian stochastic processes. The process noise w(k)

Page 47: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

46

and the measurement noise v(k) have positive semidefinite covariance matrices W and V, respectively:

)()()(E ][ T jkjwkw −= δW

)()()(E ][ T jkjvkv −= δV

0=][ T)()(E jvkw (20)

where δ(k – j) = 1 for k = j, otherwise δ(k – j) = 0. The Kalman filter estimates the state vector Δ p̂ 12(k|k) as follows:

][ )(Δ)1|(ˆΔ)()1|(ˆΔ)|(ˆΔ DFS12aw1212 kqRkkkpkkkk −−−Δ+−= pHLpp (21)

where

)1()1|1(ˆ)1|(ˆ DFS1212 −Δ+−−Δ=−Δ kqkkkk ΓpΦp (22)

is the a priori estimate of the state vector before the measurement Δpaw(k) is taken into account and L is the innovation gain, steady-state value, which can be evaluated offline before the filter operates.115,117 Values of W and V used in the model are listed in the appendix. Due to the nonzero direct transmission term R ≠ 0 in the process model (19), Δpaw(k) depends directly on the input ΔqDFS(k) as does the state estimate Δ p̂12(k | k) according to (21). The state estimate should be, however, used to set ΔqDFS(k) due to the LQ controller feedback law expressed by (15). To avoid the algebraic loop, a one-step delay block was integrated into the regulator state feedback. As can be seen in Figure 4, the output of the LQ gain block is the flow rate ΔqDFS(k), yet the Kalman filter estimates the system state Δ p̂ 12(k|k) from the process input ΔqDFS(k−1) and the process output Δpaw(k).

Bench test

A breathing simulator ASL 5000 Active Servo Lung (IngMar Medical, Pittsburgh, PA, USA) was used to test the performance of the DFS during HFO ventilation and

Page 48: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

47

for verifying the function of the controller.118-121 The objective of the experiment was to evaluate the ability of the DFS to maintain a constant mpaw during simulated spontaneous inspiration and expiration. During the bench test, the breathing simulator was connected to a SensorMedics 3100B ventilator, Figure 1b. The interconnection consisted of a Curity tracheal tube Number 8 (Kendall-Gammatron, Sampran, Thailand), an orifice flow sensor and a parabolic pneumatic resistor Rp5 (Michigan Instruments, Grand rapids, MI, USA) simulating the resistance of the real adult respiratory system. The preset flow pattern in the test lung is shown in the upper charts of Figures 5 and 6. Inspiration was simulated by a sudden increase in gas flow from the ventilator circuit into the simulator. A constant inspiratory flow rate of 30 l/min was held for 3 s. After a 3 s period of zero flow, 3 s reverse gas flow from the simulator into the ventilator circuit was introduced in order to simulate exhalation. Figure 5 presents the results of a test with the high-frequency oscillations suppressed in amplitude. Figure 6, on the other hand, presents the test results for

Figure 5 Demand flow system response without oscillations. Flow pattern between the ventilator circuit and the test lung (upper chart) and the proximal airway pressure paw measured for suppressed high-frequency oscillations (lower chart). The proximal pressure was measured with the DFS switched on (black line) and off (grey line). qaw, inspiratory and expiratory flow generated by the test lung.

Figure 6 demand flow system response with oscillations. Flow pattern between the ventilator circuit and the test lung (upper chart) and the proximal airway pressure paw measured for high-frequency oscillations at a frequency of 15 Hz and a proximal pressure amplitude of 15 cmH2O (lower chart). The proximal pressure was measured with the DFS switched on (black line) and off (grey line). qaw, inspiratory and expiratory flow generated by the test lung.

Page 49: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

48

the proximal pressure amplitude set to 15 cmH2O and oscillatory frequency set to 15 Hz. In both tests Qbf was 40 l/min and the mpaw value was 20 cmH2O. Lower charts in Figures 5 and 6 compare the proximal airway pressure waveforms of the system with the DFS in the MANUAL mode (grey line), where the DFS controller was inactive, and of a system in the DEMAND mode (black line), where the DFS was active. With the DFS controller inactive, mpaw decreased approximately by 13 cmH2O during simulated inspiration and increased approximately by 12 cmH2O during expiration. On the contrary, with the DFS in operation mpaw changed briefly at the beginning of inspiration and expiration, but was always returned to its preset value, followed by slight oscillations. The DFS trigger performance was assessed from the reaction to simulated inspiration with high-frequency oscillations suppressed. The same evaluation criteria were applied as for conventional ventilators, Figure 7: triggering delay and inspiratory delay.122 The triggering delay, defined as the time between the start of simulated breathing effort and the maximum drop in paw, was 50 ms. The inspiratory delay, defined as the time between the start of simulated breathing effort and the time when paw reached the set mpaw level again, was 115 ms.

Discussion

A unique system supporting spontaneous breathing of a patient connected to a HFO ventilator has been designed. According to the results of the completed bench test, the system is able to maintain a constant value of mpaw in the ventilator

Figure 7 Evaluation of trigger performance. Airway pressure signal paw showing the evaluation of the trigger performance. Trigger delay (TD) is defined as the time between initiation of a breath and the maximal deviation of

set mean airway pressure mpaw. Inspiratory delay (ID) is defined as the time between the start of inspiration and the time when paw is at the set mpaw level again.

Page 50: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow system design and control

49

circuit regardless of spontaneous breathing activity. This has two important benefits. A stable mpaw assures undisturbed HFO ventilator function. Next to this, from the patients’ perspective, spontaneous breathing requires less effort and should be better tolerated during HFO ventilation. The response of the DFS to simulated spontaneous breathing is presented in Figures 5 and 6. The figures show the ability of the DFS to maintain mpaw constant. The DFS trigger performance is similar to the performance of modern ventilators currently used in clinical practice. The DFS triggering delay was 50 ms and is as fast as modern mechanical ventilators.122 The inspiratory delay was 115 ms, compared to around 94 ms reported for modern ventilators. Transient pressure peaks in the paw signal develop as a reaction to sudden change in gas flow through the respiratory system. A rectangular pattern of gas flow was selected as the worst possible (and theoretical only) waveform. In a real respiratory system, the change in the gas flow is more gradual, thus the transient peaks will be significantly less expressed. Moreover, the fast pressure swings during the transient response contain high-frequency spectral components which do not propagate deep into the respiratory system.123 The fundamental requirement of the designed control system is to separate the high-frequency oscillations from spontaneous breathing in the proximal airway pressure signal. According to Figure 6, the changes in mpaw are suppressed while the high-frequency oscillations are preserved. The high-frequency oscillations are separated from the spontaneous breathing signal by a low-pass Butterworth filter with a cutoff frequency of 1.5 Hz. Energy of the high-frequency oscillations is concentrated at frequencies higher than 1.5 Hz. Nevertheless, the spontaneous breathing band is not limited to frequencies under 1.5 Hz. When the signal related to spontaneous breathing contains higher frequency components, which occur during sudden changes in paw, these transient high-frequency components cannot be fully suppressed by the DFS. The performance of the LQG controller depends on the accuracy of the model used for the description of the controlled system. Imperfections in the process model may be the reason why small swings in paw are observed when the high-frequency oscillations are switched off and the flow due to spontaneous breathing is zero, Figure 5. The present model handles both the spontaneous breathing swings and the high-frequency oscillations as white, zero-mean Gaussian stochastic processes. This assumption is inaccurate, especially for high-frequency oscillations. A more precise description of the high-frequency oscillation properties may improve the process state estimation. Furthermore, the process model does not consider inertial

Page 51: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 3

50

properties of gas which are not insignificant during HFO ventilation, when moving gas matter changes its velocity rapidly. An improved model should therefore contain inertances. Finally, the model assumes a linear relationship between mpaw and the bias flow in the ventilator circuit. Nevertheless, the resistance of the expiratory balloon valve of the HFO ventilator changes with variations in mpaw, which cannot be taken into consideration in the linear model used in the DFS.

Conclusions

The only system developed so far that effectively supports spontaneous breathing during HFO ventilation was described in this study. First of all, the DFS prevents the negative impact of the spontaneous breathing on ventilator function. The response time of the DFS system is similar to the response time of modern conventional mechanical ventilators. The effects of the DFS on work of breathing, physiologic parameters and regional lung characteristics are described in the experiments in chapter 4, 5 and 6.

Appendix

Process model parameters: R1 = 2 kPa·s/l, R2 = 0.02 kPa·s/l, R3 = 4 kPa·s/l, C1 = 1 l/kPa, C2 = 0.15 l/kPa. LQ regulator parameters: Qe = 200, Qq = 1. Kalman filter parameters: W = diag([ 105 105]), V = 10-2.

Page 52: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

Unloading work of breathing during high-frequency oscillatory ventilation: a bench study

Marc van Heerde Karel Roubik Vit Kopelent

Frans B. Plötz Dick G. Markhorst

Critical Care 2006

Page 53: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 4

52

Abstract

Introduction With the 3100B high-frequency oscillatory ventilator (SensorMedics, Yorba Linda, CA, USA), patients' spontaneous breathing efforts result in a high level of imposed work of breathing (WOB). Therefore, spontaneous breathing often has to be suppressed during high-frequency oscillatory (HFO) ventilation. A demand flow system was designed to reduce imposed WOB.

Methods An external gas flow controller (demand flow system) accommodates the ventilator fresh gas flow during spontaneous breathing simulation. A control algorithm detects breathing effort and regulates the demand flow valve. The effectiveness of this system has been evaluated in a bench test. The Campbell diagram and pressure time product (PTP) are used to quantify the imposed workload.

Results Using the demand flow system, imposed WOB is considerably reduced. The demand flow system reduces inspiratory imposed WOB by 30% to 56% and inspiratory imposed PTP by 38% to 59% compared to continuous fresh gas flow. Expiratory imposed WOB was decreased as well by 12% to 49%. In simulations of shallow to normal breathing for an adult, imposed WOB is 0.5 J/l at maximum. Fluctuations in mean airway pressure on account of spontaneous breathing are markedly reduced.

Conclusion The use of the demand flow system during HFO ventilation results in a reduction of both imposed WOB and fluctuation in mean airway pressure. The level of imposed WOB was reduced to the physiological range of WOB. Potentially, this makes maintenance of spontaneous breathing during HFO ventilation possible and easier in a clinical setting. Early initiation of HFO ventilation seems more possible with this system and the possibility of weaning of patients directly on a high-frequency oscillatory ventilator is not excluded either.

Page 54: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow during HFO ventilation: a bench study

53

Introduction

Maintenance of spontaneous breathing in mechanically ventilated patients has beneficial effects. Spontaneous breathing augments ventilation perfusion matching and cardiopulmonary function, reduces sedative requirement and shortens intensive care stay.67,69,95,97,108 High-frequency oscillatory (HFO) ventilation, at least in theory, achieves all goals of lung protective ventilation. It is a useful ventilatory mode for neonatal application,98,99 and is gaining interest in both pediatric and adult intensive care.52,101-103 Clinical trials suggest that the use of HFO ventilation at a lower severity threshold of acute respiratory distress syndrome improves outcome.58,62,101 In all adult clinical trials evaluating the efficacy of HFO ventilation, muscular paralysis was part of the study protocol.124 In larger children and adults, spontaneous breathing during HFO ventilation is currently advocated but usually not well tolerated because of patient discomfort. Consequently, early initiation of HFO ventilation has to be weighed against a high level of sedation or even muscular paralysis. In a previous bench study, we showed that spontaneous breathing of significant tidal volumes during HFO ventilation using a SensorMedics 3100B ventilator (Yorba Linda, CA, USA) leads to a high level of imposed work of breathing (WOB) and significant swings in mean airway pressure mpaw.112 The imposed WOB is the work added to the physiological WOB when breathing through a breathing apparatus. In a 3100B ventilator, this includes work to overcome resistance added by the endotracheal tube, the breathing circuit and the humidification device. The limited continuous fresh gas flow rate, that is to say, bias flow on a SensorMedics HFO ventilator, is the most important factor contributing to imposed WOB in the previous study. This high imposed WOB explains the patient discomfort.94 The normal level of physiological WOB in an adult is 0.3 to 0.6 J/l.73 The level of imposed WOB of spontaneous breathing during HFO ventilation can exceed this value by up to 400%.112 To reduce both imposed WOB and swings in mean airway pressure on account of spontaneous breathing during HFO ventilation, a demand flow system was developed. This device is capable of detecting patients' breathing efforts and subsequently adjust the fresh gas flow rate in order to reduce imposed WOB and pressure swings.

Page 55: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 4

54

Materials and methods

Ventilator principle The SensorMedics 3100B HFO ventilator used generates a mean airway pressure by a continuous fresh gas flow, with a maximum of 60 l/minute, Figure 1a. This flow passes through the patient circuit and leaves the circuit via a balloon valve. This valve is inflated to a preset pressure, leading to maintenance of a stable mean proximal airway pressure. The respiratory system is brought to its mean lung volume by this pressure, enabling oxygenation. Ventilation results from pressure oscillations generated by a loudspeaker membrane at 3 to 15 Hz superimposed upon the set mean proximal airway pressure.64

The demand flow system setup The entire demand flow system is described in chapter 3 in detail and comprises two principal components: the hardware (consisting of electronic and pneumatic parts) and the control software.

Figure 1 Scheme of the 3100B high-frequency oscillatory ventilator and the demand flow system connection. A, the basic principle of the 3100B high-frequency oscillator. B, schematic drawing of the connection of the DFS to the 3100B oscillator. HFO ventilator, high-frequency oscillatory ventilator; DFS, demand flow system paw, proximal airway pressure; Qbf, fixed bias flow rate; qDFS, bias flow rate.

Page 56: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow during HFO ventilation: a bench study

55

The hardware part of the system consists of an electronically controlled mass flow valve, proximal pressure measurement sensor with a necessary electric circuit and control and communication electronics. The valve (obtained from a commercial mechanical ventilator; AVEA, Viasys Healthcare, Yorba Linda, CA, USA) provides fresh gas flow to the ventilator circuit of the 3100B HFO ventilator, Figure 1b. A measuring board (NIDAQ 6024E, National Instruments Corporation, Austin, TX, USA) with A/D and D/A converters assures digitization of the analogue pressure signal and its transmission into a personal computer as well as conversion of control digital data from the computer into their equivalent analogue signals suitable for control of the flow valve, Figure 2. An interface connects the measuring board with the pressure sensor and a micro proportional driver that directly actuates the flow valve. paw is measured by a pressure sensor (14PC03D, Honeywell, USA) at the proximal end of the endotracheal tube. This pressure signal is preprocessed in the consequent circuit and is sent to the control computer. From this signal, patient breathing effort is detected and then a control signal is sent back to the hardware part to control the proportional valve. Fresh gas flow is regulated by patient demand. During inspiration of the patient, the flow rate is increased, and, during expiration, it is decreased. The software part of the demand flow system is responsible for analysis of the measured paw and consequent control of the flow valve. The control software is developed in a MATLAB environment (The MathWorks, Natick, USA). The measured pressure signal contains oscillations generated by the HFO ventilator and fluctuations generated by the patient breathing effort. To enable control of the

Figure 2 Schematic description of the demand flow system structure. paw, airway pressure; qDFS, bias flow rate; A/D, analog-to-digital converter; D/A digital-to-analog converter.

Page 57: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 4

56

system, the measured pressure signal, Figure 3 upper panel, is decomposed into two parts using a discreet mathematical algorithm: one component represents the ventilator pressure signal, Figure 3 middle panel, and the second part represents the patients breathing simulated with a test lung, Figure 3 lower panel. The ventilator pressure signal is a high frequency wave signal (3 to 15 Hz), in general with an asymmetrical inspiratory to expiratory time relationship. Therefore, it contains multiple higher harmonic components. The pressure signal introduced by spontaneous breathing of a patient contains lower frequency components, which allows decomposition of the measured signal into the two described parts. The patient signal is used for control of the flow valve. It modifies the delivered airflow into the ventilator circuit so that the changes in delivered airflow compensate for the pressure swings generated by the patient breathing effort. Regulation of the valve is conducted with the aim of maintaining the lowest possible deviation of the set mpaw. This is the way imposed WOB is reduced.73 Decreasing amplitude of the pressure swings when the demand flow system is in operation serves as a criterion for the control algorithm functionality. An increased airflow into the ventilator circuit during inspiration assists the patient to overcome the resistance of the ventilator circuit. It therefore reduces the breathing work required for spontaneous breathing. The system also reduces WOB during expiration because the airflow into the ventilator circuit is decreased in this phase.

Figure 3 Recording during an inspiration with continuous fresh gas flow. Top panel: pressure signal sampled at the airway opening paw. Middle panel: computed high frequency component of the pressure signal, test lung influence eliminated. Bottom panel: computed test lung induced pressure changes. The vertical line denotes the start of simulated induced inspiration. The horizontal line in the bottom panel represents set mpaw or continuous distending pressure (CDP). The curve represents fluctuation of set mpaw on account of breathing.

Page 58: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow during HFO ventilation: a bench study

57

Breathing simulation A digitally controlled test lung (high fidelity breathing simulator Active Servo Lung 5000, IngMar Medical, Pittsburgh, PA, USA) simulated spontaneous breathing. The pressure oscillations of the HFO ventilator interfered with the spontaneous breathing modes of the test lung. For this reason, the test lung was programmed as a volume pump in the 'user-defined pressure profile' mode. It generated flow patterns corresponding to spontaneous breathing flow patterns as described earlier.112 A sinusoid flow simulated inspiration of spontaneous breathing, exponentially decelerating flow expiration. The test lung was set to deliver tidal volumes Vt of approximately 340, 450 and 660 ml at a rate of 12 per minute and a series of breaths of 420 ml at a rate of 24 per minute. These settings were chosen to represent shallow and normal to deep breathing in an adult at a normal and rapid breath rate. The inspiration to expiration ratio was 1:2, as in normal breathing. Each series of breaths was preceded by a breathing pause to calculate mean proximal and lung pressures. An 8.0 mm inner diameter endotracheal tube (Rüschelit, Rüsch, Kernen, Germany) connects the test lung and the HFO ventilator. Flow through the endotracheal tube was measured with a hot-wire anemometer (Florian, Acutronic Medical Systems AG, Hirzel, Switzerland). The flow signal, the pressure signal measured at the distal end of the endotracheal tube pett and the paw signal were sampled with a sampling frequency of 100 Hz and stored for off-line analysis. Following parameters were set on the HFO ventilator: fresh gas flow 60 l/minute; mpaw 30 cmH2O; oscillatory frequency 5 Hz; and proximal pressure amplitude 80 cmH2O.

Quantification of inspiratory effort Inspiratory imposed WOB is calculated by integrating pressure measured at the distal end of the endotracheal tube pett times the volume change during inspiration:75,107

Imposed inspiratory WOB = ∫Vt,insp (CDP – mpett) dv where Vt,insp stands for inhaled tidal volume. As inspiration is active and expiration usually passive, only inspiratory imposed WOB is generally considered. Application of HFO ventilation using the SensorMedics 3100B HFO ventilator may be regarded as a super-continuous positive airway pressure system. Altering a continuous positive airway pressure device aiming at a reduction of inspiratory imposed WOB can result in an increase in expiratory imposed WOB. This may

Page 59: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 4

58

even lead to an increase in total imposed WOB.125 Therefore, expiratory imposed work of breathing was also calculated:

Imposed expiratory WOB = ∫Vt,exp (mpett – CDP) dv where Vt,exp stands for exhaled tidal volume. To enable comparison of imposed WOB in different ventilator setups, imposed WOB is often normalized, that is, related to Vt. Imposed WOB is then expressed in Joules per liter (J/l). Work per liter reflects changes in pulmonary mechanics. It is influenced by changes in resistance and compliance of the respiratory system. In a SensorMedics HFO ventilator, imposed WOB is directly related to the difference in CDP level set on the ventilator and pett; the greater the difference, the greater imposed WOB and thus patient effort. paw, Figure 4 left panels, pett and flow through the endotracheal tube were low pass filtered using a Butterworth filter, with a cut off frequency of 2 Hz. A volume signal was constructed by numerical time integration of the filtered flow signal. Subsequently, a modified Campbell diagram of each breath was plotted, Figure 4, right panels. The surface of the inspiratory part of the resulting plot represents inspiratory imposed WOB.75,107 Since imposed WOB does not reflect isometric inspiratory effort, additionally imposed pressure time product (PTP) was calculated from the filtered pett signal:80

Imposed inspiratory PTP = ∫Vt,insp (CDP - pett) dt Besides the calculation of the imposed workload, changes of paw on account of spontaneous breathing were measured using the filtered paw signal.

Results

The demand flow system reduced inspiratory imposed WOB by 30% to 56%, inspiratory imposed PTP by 38% to 59% and expiratory imposed WOB by 12% to 49% compared to using a maximum possible continuous fresh gas flow of 60 l/minute, Table 1 and Figure 4. The mpaw changed during simulated breaths, Figure 4, left panels. With continuous fresh gas flow mpaw fluctuated on average between -17 cmH2O below the CDP during inspiration to +11 cmH2O above the CDP during expiration, Table 1. With the demand flow system, these fluctuations

Page 60: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow during HFO ventilation: a bench study

59

were smaller: -9 to +8 cmH2O. According to the oscillator manual, as a safety measure, proximal pressure alarm limits should be set 3 cmH2O below and above the CDP. With continuous fresh gas flow, mpaw was outside these safety limits 25% of the time during simulated spontaneous breathing during inspiration and 33% during expiration. With the demand flow system it was 11% of time during inspiration and 12% during expiration. Using the continuous fresh gas flow ventilator, alarms sounded constantly during all simulations. Using the demand flow system this only occurred at a breathing simulation with a Vt of 420 ml 24 per minute.

Discussion

Spontaneous breathing during HFO ventilation results in significant fluctuations of mean airway pressure. We recently described that this results in a high level of inspiratory WOB.94 For the operator, the pressure fluctuations hamper titration of the desired set airway pressure while the pressure safety alarm limits may be exceeded frequently. This experiment demonstrates that, in a bench test, the use of the demand flow system decreases imposed WOB significantly. It also limits breathing induced fluctuation of proximal airway pressure and time where proximal pressure exceeds safety alarm limits during breathing. Although the demand flow algorithm was primarily designed to reduce inspiratory WOB by increasing fresh gas flow to meet patient demand, this did not lead to

Table 1 Summary of test results with continuous flow and with the demand flow system.

Simulated breaths Δ mpaw (cmH2O) WOBi (J/l) PTPi (cm/H2O/s) WOBe (J/l)

Vt (ml) Rate (/min) CF DFS CF DFS CF DFS CF DFS

330 12 - 10 + 6 - 4 + 4 0.85 0.37 12 5.1 0.51 0.26

450 12 - 13 + 9 - 6 + 5 1.2 0.50 17 6.8 0.69 0.38

66 12 - 20 + 14 - 9 + 7 1.8 0.78 26 11 1.1 0.61

420 24 - 24 + 16 - 15 + 17 2.2 1.5 16 10 1.4 1.2

Simulated breaths: Vt; tidal volume, rate; respiratory rate. Δ mpaw; deviation of mean proximal airway pressure from set mpaw during inspiration (negative values) and expiration (positive values). WOBi; inspiratory imposed work of breathing. WOBe; expiratory imposed work of breathing PTPi; inspiratory imposed pressure time product.

Page 61: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 4

60

increase in pressures during expiration. Expiratory imposed WOB was reduced as well. The effectiveness of the demand flow system in decreasing the imposed WOB was less marked when breath rate increased from 12 to 24 per minute. In pediatric patients, but also in some adults, with severe acute respiratory distress syndrome, high breathing rates at small tidal volumes are clinically observed. The effectiveness of the demand flow system needs, therefore, to be tested in vivo. The optimal workload for critically ill patients is unclear. It depends on energy and muscular reserve. Research focuses mainly on WOB in the weaning phase.81,82 A WOB level in the physiological range, approximately 0.5 J/l in adults, seems to correspond with an optimal workload. Full unloading, for instance reducing the WOB to zero, induces loss of respiratory muscles. Excessive respiratory muscle loading may cause muscle fatigue and weaning failure.82 This workload of 0.5 J/l seems to be optimal not only during weaning, but also in the acute phase of respiratory failure.83,108 Compared to the WOB of a healthy adult (0.3 to 0.6 J/l), imposed WOB is high if spontaneous breathing is simulated during HFO ventilation using continuous fresh gas flow. Using the demand flow system,

Figure 4 Pressure recordings of proximal airway pressure and modified Campbell diagrams during simulated spontaneous breathing. Pressure recordings of proximal airway pressure (paw) (left panels) and modified Campbell diagrams (right panels) during simulated spontaneous breathing. Upper panels show recordings with continuous fresh gas flow; bottom panels show recordings with the demand flow system. The left panels depict paw variation during two subsequent breaths. Thin lines represent unfiltered pressure signals and thick lines represent filtered paw. Note the reduced changes in both unfiltered and filtered signal with the demand flow system (imposed pressure time product 17 cmH2O/s versus 6.8 cmH2O/s). Lines in the right panels represent mean lung pressure pett. Note the reduced surface area in the lower right panel. Imposed work of breathing is 1.2 J/l without the demand flow system versus 0.5 J/l in the lower right panel with the demand flow system. pett, pressure at end of the tracheal tube.

Page 62: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Demand flow during HFO ventilation: a bench study

61

imposed WOB is considerably reduced. In simulations of shallow to normal breathing for an adult, imposed WOB was 0.5 J/l at maximum. The SensorMedics 3100A HFO ventilator was originally designed for neonatal application. The 3100B ventilator was thereafter designed for oscillating patients weighing more than 35 kg. Although small neonatal patients can breathe comfortably on their HFO ventilation circuit, paralysis has been felt necessary in most larger patients in whom spontaneous breathing imposes a high level of WOB and triggers numerous alarms, with resultant loss of the benefits of maintaining some element of spontaneous respiration during ventilator support. This is a significant disadvantage that needs rectification.99,112 The demand flow system seems capable of achieving this.

Limitations of the study

In this in vitro study we aimed to choose realistic test conditions. The limitations of the bench test model were discussed in chapter 2.94 In addition, patient ventilator interaction cannot completely be simulated in a bench test. Whether the demand flow system is capable of providing comfortable ventilation synchrony, for instance, needs to be tested in vivo.

Conclusions

A novel demand flow system has been designed that is capable of automatic regulation of HFO ventilator fresh gas flow adjusted to patient need during simulated spontaneous breathing. This results in a considerable reduction of imposed WOB and reduction of swings in mean airway pressure. The amount of reduction in imposed WOB is promising. Potentially, this may lead to a reduced use of sedatives and muscular paralysis in larger patients during HFO ventilation. Early initiation of HFO ventilation seems more possible with this system and the possibility of weaning of patients directly on a HFO ventilator is not excluded either.

Page 63: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging
Page 64: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

Demand flow facilitates spontaneous breathing during high-frequency oscillatory ventilation in a pig model

Marc van Heerde Karel Roubik Vit Kopelent

Frans B. Plötz Dick G. Markhorst

Critical Care Medicine 2009

Page 65: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 5

64

Abstract

Objective: Maintenance breathing is advocated in mechanical ventilation, which is difficult for the high-frequency oscillatory (HFO) ventilation. To facilitate spontaneous breathing during HFO ventilation, a demand flow system (DFS) was designed. The aim of the present study was to evaluate the system. Design: Animal experiment. Setting: University animal laboratory. Subjects: Eight pigs (47-64 kg). Interventions: Lung injury was induced by lung lavage with normal saline. After spontaneous breathing was restored HFO ventilation was applied, in runs of 30 minutes, with continuous fresh gas flow (CF) or the DFS operated in two different setups. Pressure to regulate the DFS was sampled directly at the Y-piece of the ventilator circuit (DFS) or between the endotracheal tube and measurement equipment at the proximal end of the endotracheal tube (DFSPROX). In the end, animals were paralyzed. Breathing pattern, work of breathing, and gas exchange were evaluated. Measurements and main results: HFO ventilation with demand flow decreased breathing frequency and increased tidal volume compared with CF. Comparing HFO modes CF, DFS, and DFSPROX, total pressure time product (PTP) was 66 cmH2O/s/min (interquartile range 59-74), 64 cmH2O/s/min (50-72), and 51 cmH2O/s/min (41-63). Ventilator PTP was 36 cmH2O/s/min (32-42), 8.6 cmH2O/s/min (7.4-10), and 1 cmH2O/s/min (-1.0 to 2.8). Oxygenation, evaluated by PaO2, was preserved when spontaneous breathing was maintained and deteriorated when pigs were paralyzed. Ventilation, evaluated by PaCO2, improved with demand flow. PaCO2 increased when using continuous flow and during muscular paralysis. Conclusions: In moderately lung-injured anesthetized pigs during HFO ventilation, demand flow facilitated spontaneous breathing and augmented gas exchange. Demand flow decreased total breathing effort as quantified by PTP. Imposed work caused by the HFO ventilator appeared totally reduced by demand flow.

Page 66: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: an animal study

65

Introduction

Acute lung injury and acute respiratory distress syndrome affect the lung heterogeneously. Both experimental models and clinical studies emphasize the positive effect of spontaneous breathing during mechanical ventilation (MV) on distribution of inflation and ventilation in the diseased lung. Spontaneous breathing improves oxygenation, lowers need for sedatives, improves hemodynamics, and reduces duration of MV and intensive care stay.66,67,70,96 An open lung approach, as described by Froese42 and Lachmann,44 reverses atelectasis, avoids overdistension of open lung units, and protects the injured lung from further harm. High-frequency oscillatory (HFO) ventilation, with an open lung strategy, is, in theory, a modality that can achieve optimal lung protection. Early application of HFO ventilation seems to give the optimal lung protection.50,101,124 In HFO ventilation, more conventional respiratory rates (RRs) and tidal volumes Vt are not needed to achieve adequate gas exchange.64 Preservation of spontaneous breathing during MV was not yet an issue during the development of the HFO ventilator (SensorMedics, 3100 A/B, Yorba Linda, CA) in the 1970s and 1980s. To have patients spontaneously breathe was, therefore, not the focus of the design of the HFO ventilator. As a result, spontaneous breathing during HFO ventilation is not well tolerated in large pediatric and adult patients, which is caused by a high imposed workload added by the HFO ventilator.112 The use of HFO ventilation for life-sustaining gas exchange is counterbalanced by the need for heavy sedation and possible muscular paralysis.94,112 Furthermore, weaning from the HFO ventilator may be prolonged due to sedative and paralytic use. In an HFO ventilator, the fixed continuous fresh gas flow (CF) is the most important factor defining the imposed work of breathing (WOB). A demand flow system (DFS) was developed to advance to better HFO ventilation strategies that incorporate spontaneous breathing of a patient. In a bench study, we already demonstrated that the imposed WOB decreased considerably when demand flow was used instead of CF.126 The aim of this study was to evaluate the influence of our DFS with HFO ventilation on different components of breathing effort, on respiratory variables, and on gas exchange in a pig model of acute lung injury.

Page 67: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 5

66

Materials and methods

The study was approved by the Animal Welfare Committee of the VU University Medical Center. Eight Dalland pigs (body weight range, 47–64 kg) were used.

Animal Preparation Anesthesia: Anesthesia was induced with intramuscular injection of 0.5 mg atropine, 0.5 mg/kg midazolam, and 10 mg/kg ketamine. After induction, an ear vein was cannulated and propofol 3 mg/kg was injected before endotracheal intubation with a cuffed tube (inner diameter 8 mm). Anesthesia was maintained with continuous infusion of propofol 4 mg/kg/hr and remifentanil 0.4 μg/kg/min during instrumentation and lung lavage. To allow spontaneous breathing, propofol dosage was lowered to 2 mg/kg/hr, and that of remifentanil to 0.05–0.1 μg/kg/min. When necessary according to the experimental protocol, spontaneous breathing was suppressed using pancuronium bromide 0.3 mg/kg/hr. At the end, animals were killed using sodium pentobarbital. Surgical preparation: During instrumentation, lung lavage, and the stabilization period, animals were ventilated with a Servo 900C ventilator (Maquet Critical Care AB, Solna, Sweden) in a volume-controlled mode with the following settings and then adjusted to maintain normocapnia (PaCO2 38–45 mmHg): RR 20 /min, inspiratory pause time 0.6 seconds, positive end-expiratory pressure 5 cmH2O, inspiration to expiration ratio 1:2, FiO2 1.0, initial Vt 10 ml/kg. Animals were placed in supine position on a heated table. Temperature was kept in the normal range (38–39°C) using a heating pad. The left femoral artery was cannulated to measure arterial blood pressure and to sample blood. A Paratrend 7 continuous intravascular blood gas monitor (Biomedical Sensors, High Wycombe, United Kingdom) was inserted at the left femoral artery. A triple lumen pulmonary arterial catheter was inserted to monitor pulmonary arterial and central venous pressures and to sample mixed venous blood. A separate catheter was inserted into the superior vena cava to infuse fluids and anesthetics. Surfactant depletion: Surfactant deficiency was induced by a repeated whole lung lavage. Normal saline 30–40 ml/kg of 37°C was instilled in the lungs at a pressure of 50 cmH2O and then directly removed by drainage. The lavage was repeated after 1 hour.127,128 HFO ventilator: A SensorMedics 3100B HFO ventilator (SensorMedics, Yorba Linda, CA) was used. In the HFO ventilator, mean airway pressure mpaw is

Page 68: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: an animal study

67

maintained by two mechanisms: setting of a CF and setting of the resistance of the expiratory balloon valve. A patients’ spontaneous breathing during HFO ventilation generates changes in mpaw. The changes in mpaw determine directly the workload for the patient; the higher the changes, the higher the workload.129 The standard HFO ventilator cannot compensate for changes in mpaw caused by spontaneous breathing. To solve the problem, the 3100B HFO ventilator was equipped with a custom-made DFS. A detailed description of the DFS is given in chapter 3.

Measurements and Samples Data acquisition: The experimental setup is depicted in Figure 1A. Flow was measured at the proximal end of the endotracheal tube using a hot-wire anemometer (Florian, Acutronic Medical Systems AG, Hirzel, Switzerland). For

Figure 1 Experimental setup and signal processing. A, Pressure to regulate the demand flow system (DFS) is sampled at two different sites: Prox, between the endotracheal tube and measuring equipment proximal end of the endotracheal tube or Y-piece, between ventilator circuit and measuring equipment. pes, esophageal pressure; ptrach, tracheal pressure; paw, pressure at Y-piece; HFO ventilator, high-frequency oscillatory ventilator; Vt, tidal volume. B, left: computed tidal volume of spontaneous breathing. •, start of inspiration; *, end of inspiration. Right: restoration of tidal breathing to same start volume.

Page 69: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 5

68

measurement of the tracheal pressure ptrach with the respiratory monitor, an air-filled 5F catheter was introduced into the endotracheal tube, its tip located at the distal end of the tube. An esophageal balloon catheter (SmartCath Catheter 8F, Viasys Healthcare, Palm Springs, CA) was placed for measurement of esophageal pressure pes to approximate pleural pressure. pes was sampled using an analog pressure sensor (40PC, Honeywell, Morristown, NJ). Validation of proper placement of the esophageal catheter was done using the occlusion test.130 The pressure at the Y-piece in the ventilator circuit paw was sampled using the unfiltered electronic signal from the internal pressure sensor of the HFO ventilator. Pressure sensors were calibrated using a water column. Flow and pressure signals were recorded at 100 Hz and stored on a laptop computer for off-line analysis. Data processing: A MATLAB environment was used for data processing (The MathWorks, Natick, MA). In each animal, 5-minute segments of air flow and pressure signals were recorded for different HFO ventilation modes. For evaluation, 2-minute segments with a regular breathing pattern were studied from the 5-minute recordings. To eliminate HFO ventilator oscillations, the recorded signals were low-pass filtered using a seventh-order Butterworth filter with a cutoff frequency of 2.5 Hz. The filtered flow signal represented flow changes caused by spontaneous breathing of the pigs. The initial volume at the start of each breath changed in time, Figure 1B left. Using linear interpolation, a zero baseline was created, Figure 1B right. Data evaluation: From the integrated filtered flow signals, breathing pattern and minute ventilation were determined for each individual breath and averaged over a 2-minute period. To evaluate the influence of the DFS with HFO ventilation on different components of breathing effort, inspiratory pressure time product (PTP) was evaluated.131,132 Total PTP was calculated as the area between elastic recoil pressure of the chest wall pescw and pes.132 Static chest wall compliance Ccw was measured to calculate pescw. Ccw was determined during muscular paralysis, after the second lung lavage, by inflating the lungs with a volume of 1.5 liter with a syringe (3.0 liter Calibrated Syringe, Hans Rudolph, Kansas City, MO). Ccw was calculated as the Vt divided by the difference in inspiratory and expiratory pes at points of zero flow. Total imposed and ventilator PTP were computed as the area between set mpaw and ptrach and paw, respectively. Hemodynamic and respiratory variables: Arterial and mixed venous blood samples were analyzed with ABL505 and OSM3 hemoximeters (Radiometer, Copenhagen, Denmark). Continuous arterial blood gas analysis was conducted by the

Page 70: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: an animal study

69

Paratrend 7. Physiologic shunt fraction (Qs/Qt) and respiratory indices were calculated according to standard formulas.133

Experimental Protocol Figure 2 shows the study design. After a 30-minute stabilization period on conventional ventilation, HFO ventilation was initiated. Initial settings were as follows: continuous distending pressure (CDP) 20 cmH2O, proximal pressure amplitude (ΔP) was set to maintain normocapnia (38–45 mmHg), oscillatory frequency 5 Hz, inspiration/expiration ratio 1:2, fresh gas flow 40 l/min, and FiO2 1.0. The order of the three different HFO modalities with spontaneous breathing was randomly determined: 1, HFO ventilation with a CF of 20 l/min (CF); 2, HFO ventilation with demand flow, where pressure to regulate the DFS was sampled directly at the Y-piece of the ventilator circuit (DFS); and 3, HFO ventilation with demand flow, where pressure to regulate the DFS was sampled at the proximal end of the endotracheal tube (DFSPROX). The two different pressure pick-up points were necessary to evaluate the influence of the measuring equipment, Figure 1, on the performance of the DFS. In a fourth step, all animals studied were totally paralyzed. To standardize lung volume at the start of each HFO ventilation mode, a recruitment maneuver was performed.44,134 Initially, mpaw on HFO ventilation was increased to 30 cmH2O for 5 minutes. mpaw Was lowered to 25 cmH2O when heart rate or blood pressure was unstable for >2 minutes. mpaw Was then lowered until the animals started breathing in a regular pattern. Measurements were done in the last 5 minutes of every 30 minutes of different HFO ventilation modalities.

Statistical Analysis Data are expressed as median and 25th to 75th interquartile range. Parameter comparison for different HFO ventilation modes was performed using repeated-measures analysis of variance with Bonferroni post hoc testing. In all analysis,

Time random order

conventional ventilation high-frequency oscillatory ventilation

instrumentation lavage stabilization REC CF REC DFS REC DFSPROX REC paralyzed

Figure 2 Study design. Continuous flow, CF; demand flow with (1) pressure sampled at the Y-piece (DFS) and (2) pressure sampled at the proximal end of the endotracheal tube (DFSPROX).

Page 71: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 5

70

a p value of < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS 15 for Windows (SPSS, Chicago, IL).

Results

All eight animals completed the entire experimental protocol.

Respiratory Variables The respiratory variables obtained during different HFO modalities are summarized in Table 1. Minute ventilation of spontaneous breathing (MV) was equal in all three HFO ventilation modes. Vt and RR significantly differed between HFO ventilation modes. Vt is lowest and RR is highest in the CF modality; in the DFSPROX mode, this was the reverse. Vt per kg increased from 4.3 (4.2–4.7) in the CF mode to 8.1 ml/kg (6.8–9.1) in the DFSPROX mode. RR was inversely related to Vt. An increase of Vt was paralleled by an increase of inspiratory time Ti.

Breathing Effort PTP per minute was significantly lower for all different components in the DFSPROX mode compared with the CF and DFS mode. Total imposed and ventilator PTP per minute in the DFS mode were lower compared with the CF mode. No active expiration was observed. The effectiveness of the DFS system

Table 1 Respiratory variables during different high-frequency oscillatory ventilation modes

CF DFS DFSPROX

Respiratory rate (/min) 8.5 (7.8 - 9.4) 7.6 (6.4 - 9.9) a 5.0 (4.1 - 5.8) a,b

Tidal volume (l) 0.23 (0.22 - 0.25) 0.28 (0.22 - 0.34) a 0.43 (0.36 - 0.48) a,b

Tidal volume (ml/kg) 4.3 (4.2 - 4.7) 5.1 (4.2 - 6.4) a 8.1 (6.8 - 9.1) a,b

Minute ventilation (l/min) 2.1 (1.9 - 2.2) 2.2 (1.9 - 2.4) 1.9 (1.7 - 2.4)

Inspiratory time (s) 1.5 (1.5 - 1.8) 1.7 (1.4 - 2.0) a 2.1 (1.9 - 2.2) a,b

Pressure Time Product per minute (cmH2O/s/min)

Total 66 (59 - 74) 64 (50 - 72) 51 (41 - 63) a,b

Total imposed 67 (58 - 78) 51 (41 - 57) a 36 (28 - 43) a,b

Ventilator 36 (32 - 42) 8.6 (7.4 - 10) a 1.0 (-1.0 - 2.8) a,b

a p < 0.05; CF vs. DFS or DFSPROX; b p < 0.05 DFS vs. DFSPROX. CF, continuous fresh gas flow; DFS, demand flow system; DFSPROX, Demand flow system with pressure sampled at proximal end of endotracheal tube.

Page 72: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: an animal study

71

to lower breathing effort for different Vt values of spontaneous breathing is depicted in Figure 3A–C. For a given Vt, different components of PTP per cycle were highest when CF mode was used. Compared with CF total PTP per cycle decreased 10% to 39% using DFS, and decreased 39% to 48% using DFSPROX. In Figure 4, the effect of the DFS on mpaw and Vt is illustrated. The DFS is able to reduce changes in mpaw on account of spontaneous breathing. When pressure to regulate the DFS was sampled directly at the proximal end of the endotracheal tube, pressure support was added during inspiration in the example. During expiration, paw is lower than set mpaw. The fact that demand flow generates a higher Vt is also

Figure 3: Correlation between tidal volume and total, total imposed, and ventilator pressure–time product per breathing cycle. Correlation between tidal volume and (A) total, (B) total imposed and (C) ventilator pressure time product (PTP) per breathing cycle. Continuous flow (▬), demand flow with pressure sampled at the Y-piece (─) and at the proximal end of the endotracheal tube (- -). Lines represent second order polynomial regression curves.

Page 73: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 5

72

demonstrated. All three different breaths have approximately equal inspiratory PTP per cycle. With increase in support of the DFS, the flow and thus Vt increases clearly.

Gas Exchange Results for gas exchange are shown in Table 2. Oxygenation improved in all HFO ventilation modes when spontaneous breathing was maintained. When pigs were paralyzed, PaO2 decreased remarkably during the 30-minute period. PaCO2 decreased in the DFS and DFSPROX mode. PaCO2 increased significantly using CF or during muscular paralysis.

Discussion

Our animal study demonstrates that demand flow facilitates spontaneous breathing during HFO ventilation by lowering WOB. The DFS was able to effectively minimize work imposed by the ventilator. The amount of support during spontaneous breathing can be influenced by changing the pressure sampling site to

Table 2 Physiologic variables during different high-frequency oscillatory ventilation modes.

spontaneously breathing paralyzed

CF DFS DFSPROX PAR

PaO2 (mmHg) 493 (455-502) 473 (420-502) 501 (493-554) 451 (420-489)

ΔPaO2 (mmHg) 19.9 (19.6 - 37) 31 (-5.5 to 84) 19.3 (5.5 - 34) -140 (-225 to -53) a

PaCO2 (mmHg) 53 (52-54) 49 (46-52) 52 (48-53) 70 (63-72)

ΔPaCO2 (mmHg) 6.3 (-3.0 - 18) -5.7 (-4.3 to -7.4) b -4.9 (-3.8 to -7.0) b 9.9 (3.6-22) c

Aa-DO2 (mmHg) 150 (145-192) 174 (145 - 198) 136 (95-148) 168 (125-216)

Qs/Qt 0.06 (0.04-0.08) 0.1 (0.06-0.3) 0.01 (0.002 – 0.3) 0.3 (0.29-0.3)

mpaw (cmH2O) 10 (9.9-11) 10 (8.7-11) 10 (9.9 – 11) 10 (9.9-11)

Oxygenation index 2.1 (2.0-2.2) 2.0 (1.9 – 2.1) 2.0 (1.8-2.1) 2.1 (2.0-2.5)

a p < 0.05 PAR vs. CF, DFS and DFSPROX; b CF vs. DFS and DFSPROX; c PAR vs. DFS and DFSPOX. CF, continuous fresh gas flow; DFS, demand flow system; DFSPROX demand flow with pressure sampled at proximal end of endotracheal tube; PAR, muscular paralysis. ΔpaO2 and ΔpaCO2, difference in PaO2 and PaCO2 at the beginning and end of 30 minutes HFO ventilation; Aa-DO2, alveolar-arterial pO2 difference; Qs/Qt, physiologic shunt fraction; mpaw, mean airway pressure; OI, oxygenation index.

Page 74: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: an animal study

73

regulate the DFS. Even additional pressure support can be generated to overcome WOB imposed by the endotracheal tube.

Respiratory variables The spontaneous breathing pattern changed when different modalities of HFO ventilation were compared. In HFO ventilation with CF, Vt was lowest and breathing frequency highest. In the DFSPROX mode, Vt was highest and breathing frequency lowest. These results are consistent with the theory of minimal work. The breathing frequency adopted by the animal to achieve a target minute ventilation represents a strategy to minimize inspiratory effort.135 The lower RRs observed with the DFS in use was caused by the effective reduction in WOB by the DFS. The Vt in the DFSPROX mode increased to 8.1 ml/kg (6.8–9.1). These Vt

Figure 4: Illustration of pressures and flow during different high-frequency ventilation modes. Paw, airway pressure; ptrach, tracheal pressure; mpaw, set mean airway pressure on the high-frequency oscillatory ventilator; pescw, elastic recoil pressure of the chest wall; pes, esophageal pressure; et-tube, endotracheal tube; insp., inspiration. Depicted are three different breaths: 1, left with continuous fresh gas flow (CF); 2, middle with demand flow where pressure was sampled at the Y piece (demand flow system [DFS]); and 3, right with demand flow and pressure sampled at the proximal end of the endotracheal tube (DFSPROX). Note that due to the demand flow system, the swings in paw decrease comparing CF and DFS. When pressure was sampled at the proximal end of the endotracheal tube, during inspiration there was some pressure support, as paw becomes higher than set mpaw. There was no pressure overshoot as ptrach is lower than mpaw during inspiration. During expiration, paw is lower than set mpaw to facilitate expiration. The area PTP represents the total inspiratory pressure–time product. Note that with almost equal PTP for all three breaths, the DFSPROX generates the highest flow and thus the highest Vt.

Page 75: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 5

74

values are not exceptionally high and are not expected to cause additional lung injury.136 The lung lavage model used did not cause widespread atelectasis and consolidation in the animal lungs, as concluded from the data on gas exchange. Whether an increase in Vt also improved the distribution of ventilation is impossible to say based on the results and is an area of further research.

Breathing Effort Total, total imposed, and ventilator PTP per minute were lowest in the DFSPROX mode. The DFS is effective in reducing PTP per cycle for different Vt values, Figure 3. Comparison of the results with earlier bench tests show differences in the amount of imposed PTP.126 In the animal study, the amount of imposed workload was substantially lower compared with earlier bench test results. The difference in imposed breathing work can be explained by a lower breathing frequency of the animals compared with the breathing frequencies simulated in the bench tests.

Gas Exchange The results for gas exchange show a benefit of spontaneous breathing during HFO ventilation. PaO2 improved in all HFO ventilation modes when spontaneous breathing was maintained. Muscular paralysis had a deteriorating effect on PaO2. Accordingly, there is a trend of worsening of physiologic shunt fraction when animals were paralyzed. The trend did not reach statistical significance. Use of the demand flow during HFO ventilation improved ventilation as is illustrated by a decrease in PaCO2 in the DFS and DFSPROX mode. PaCO2 increased significantly using CF or during muscular paralysis. To observe the effect of spontaneous breathing and muscular paralysis on gas exchange, HFO ventilator settings were not changed during the 30 minutes of each HFO mode. The pigs only had mild lung injury. Therefore by adjustment of HFO ventilator settings, the derangement in gas exchange would have been easily corrected. The results on gas exchange, however, indicate that, when spontaneous breathing is maintained, adequate gas exchange may be achieved at a lower mpaw.

Synchronization MV assumes the WOB, improves gas exchange, and unloads the respiratory muscles, all of which require good synchronization between the patient and the ventilator.137 In conventional MV, synchrony is evaluated by analysis of triggering and flow and pressure recordings. The DFS has no flow or pressure trigger comparable with conventional mechanical ventilators. Simplified, the DFS

Page 76: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

WOB during HFO ventilation: an animal study

75

responds to changes in mpaw caused by spontaneous breathing with the following approach: The DFS algorithm predicts an oscillatory paw signal. Because of spontaneous breathing, the actual paw signal will change. When the DFS detects a difference between the predicted and actual paw, it adjusts the fresh gas flow to maintain a stable mpaw. Furthermore, the DFS algorithm is adaptive over time. The response time of the DFS, to react on changes in mpaw, will therefore change in time. During the experiment flow over time generated by the DFS was not recorded. Therefore synchrony between the DFS and animal’s spontaneous breathing pattern could not be evaluated. The animals at least showed no signs of air hunger but were, of course, sedated.

Limitations

Our study has some limitations. In a clinical setting, HFO ventilation is applied in severe acute lung injury/acute respiratory distress syndrome. We initially thought of using a model of severe lung injury, which necessitates high mpaw to keep oxygenation in acceptable ranges; high mpaw, however, induces apnea in pigs (H. Wrigge, Department of Anesthesiology and Intensive Care Medicine of the University of Bonn, personal communication). Therefore, to test our DFS we could apply only mild lung injury. Our results are, thus, restricted to low mpaw and to mild lung injury. Additionally, we could not test high RR, which is regrettable, as our DFS was less effective at higher than at lower RR in bench tests.126 Flow and pressure sensors, necessary for data acquisition added considerable resistance to the ventilator circuit. The additional resistance has a negative effect on the DFS performance. By regulating the DFS at different pressure sampling sites, the additional resistance was partly overcome.

Conclusions

During HFO ventilation in surfactant-depleted pigs, DFS facilitates spontaneous breathing, reduces breathing effort, and improves gas exchange. Demand flow may prove to be a valuable tool in reducing the threshold for early application of HFO ventilation. In addition, demand flow may augment weaning during HFO ventilation. The development of a flow demand system for a HFO ventilator is an important step toward further improvement of HFO ventilation strategies that incorporate spontaneous breathing of a patient.

Page 77: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 5

76

Further studies are needed to elucidate DFS performance at higher mean airway pressures and higher respiratory rates seen in clinical use of HFO ventilation. Furthermore, patient and DFS synchrony needs to be evaluated.

Page 78: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

Effect of spontaneous breathing during high-frequency oscillatory ventilation on regional lung characteristics in experimental lung injury

Marc van Heerde Karel Roubik Vit Kopelent

Martin C.K. Kneyber Dick G. Markhorst

Acta Anaesthesiologica Scandinavica 2010

Page 79: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 6

78

Abstract

Background Maintenance of spontaneous breathing is advocated in mechanical ventilation. This study evaluates the effect of spontaneous breathing on regional lung characteristics during high-frequency oscillatory (HFO) ventilation in an animal model of mild lung injury.

Methods Lung injury was induced by lavage with normal saline in eight pigs (weight range 47–64 kg). HFO ventilation was applied, in runs of 30 minutes on paralyzed animals or on spontaneous breathing animals with a continuous fresh gas flow (CF) or a custom-made demand flow (DF) system. Electrical impedance tomography (EIT) was used to assess lung aeration and ventilation and the occurrence of hyperinflation.

Results End expiratory lung volume (EELV) decreased in all different HFO modalities. HFO, with spontaneous breathing maintained, showed preservation in lung volume in the dependent lung regions compared with paralyzed conditions. Comparing DF with paralyzed conditions, the center of ventilation was located at 50% and 51% (median, left and right lung) from anterior to posterior and at 45% and 46% respectively, p < 0.05. Polynomial coefficients using a continuous flow were -0.02 (range -0.35 to 0.32) and -0.01 (-0.17 to 0.23) for CF and DF, respectively, p = 0.01.

Conclusions This animal study demonstrates that spontaneous breathing during HFO ventilation preserves lung volume, and when combined with DF, improves ventilation of the dependent lung areas. No significant hyperinflation occurred on account of spontaneous breathing. These results underline the importance of maintaining spontaneous breathing during HFO ventilation and support efforts to optimize HFO ventilators to facilitate patients’ spontaneous breathing.

Page 80: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Regional lung characteristics during HFO ventilation: an animal study

79

Introduction

Once the mechanisms responsible for ventilator-induced lung injury (VILI) were elucidated, ventilator strategies were sought that could protect the already injured lung from additional harm.6,42 Lung-protective ventilation strategies aim at the reversal of atelectasis and avoidance of hyperinflation as well as cyclic opening and closing of lung units during tidal ventilation. High-frequency oscillatory (HFO) ventilation is, at least in theory, a ventilation modality that can achieve optimal lung protection. Several animal studies show that HFO ventilation reduces VILI.51,138 Both experimental and clinical studies emphasize the positive effect of spontaneous breathing during mechanical ventilation on the distribution of inflation and ventilation in the diseased lung. Spontaneous breathing improves oxygenation, lowers the need for sedatives, improves hemodynamics and reduces the duration of mechanical ventilation and intensive care stay.67,96,139,140 In HFO ventilation, more conventional respiratory rates (RR) and tidal volumes, as in spontaneous breathing efforts, are not needed to achieve adequate gas exchange.64 Vigorous spontaneous breathing during HFO ventilation in large patients may in fact cause swings in set mean airway pressure mpaw that activate alarms, interrupt oscillations and produce significant desaturations. Initial adult HFO ventilation trials recommended muscular paralysis for this reason.141,142 The current HFO ventilators’ design (3100A/B, SensorMedics, Yorba Linda, CA) with a fixed fresh gas rate makes it difficult to breathe during HFO ventilation.112 In chapter 5, we report on this animal experiment focused on work of breathing during HFO ventilation.143 We demonstrated that demand flow (DF), instead of a continuous fresh gas flow (CF), facilitated spontaneous breathing. The focus of this part of the experiment was to evaluate the effect of spontaneous breathing during HFO ventilation on lung aeration and ventilation in a porcine model of moderate lung injury with the use of electrical impedance tomography (EIT). EIT is a noninvasive technique for pulmonary imaging. EIT provides meaningful dynamic information on pulmonary conditions. EIT is able to accurately describe both global and regional lung volume changes over time during mechanical ventilation.144 Compared with electron beam computed tomography, as an established method to assess changes in local air content, simultaneous EIT measurements correlate quite closely.144-146

Page 81: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 6

80

Materials and methods

Animal model preparation The study was approved by the local Animal Welfare Committee. Eight Dalland pigs 53 ± 6.5 kg (mean ± SD) were used. Anesthesia was induced with an intramuscular injection of 0.5 mg atropine, 0.5 g/kg midazolam and 10 mg/kg ketamine. After induction, propofol 3 mg/kg was injected intravenously before endotracheal intubation with a cuffed tube (inner diameter 8 mm). Anesthesia was maintained with continuous infusions of propofol 4 mg/kg/h and remifentanil 0.4 μg/kg/min during instrumentation and lung lavage. To allow spontaneous breathing, propofol dosage was reduced to 2 mg/kg/h, and that of remifentanil to 0.05–0.1 μg/kg/min. Spontaneous breathing was suppressed using pancuronium bromide 0.3 mg/kg/h during instrumentation, lung lavage and when necessary according to the experimental protocol. At the end, animals were euthanized using sodium pentobarbital. During instrumentation, lung lavage and the stabilization period, animals were ventilated using a Servo 900C ventilator (Maquet Critical Care AB, Solna, Sweden) in the volume-controlled mode: RR 20/min, inspiratory pause time 0.6 s, positive end expiratory pressure 5 cmH2O, inspiration to expiration ratio 1:2, FiO2 1.0, initial tidal volume Vt 10 ml/kg and then adjusted to maintain normocapnia (PaCO2 38–45 mmHg). Animals were placed in a supine position on a heated table. Temperature was maintained in the normal range (38–390C) using a heating pad. The left femoral artery was cannulated to measure blood pressure and sample blood. A pulmonary arterial catheter was inserted to sample mixed venous blood. A separate catheter was inserted into the superior vena cava to infuse fluids and anesthetics. Flow was measured at the proximal end of the endotracheal tube using a hot-wire anemometer (Florian, Acutronic Medical Systems AG, Hirzel, Switzerland). The pressure at the Y-piece in the ventilator circuit paw was sampled using the electronic signal from the internal pressure sensor of the HFO ventilator. The pressure sensor signal was calibrated using a water column. Flow and pressure signals were recorded at 100 Hz and stored on a laptop computer for off-line analysis. Surfactant deficiency was induced by repeated whole lung lavage. Normal saline 30–40 ml/kg of 37°C was instilled in the lungs at a pressure of 50 cmH2O and then directly removed by drainage. The lavage was repeated after one hour.127,128

Page 82: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Regional lung characteristics during HFO ventilation: an animal study

81

HFO ventilator HFO ventilation was applied using the 3100B HFO ventilator (SensorMedics). The HFO ventilator was used in a standard configuration with a CF. In order to facilitate spontaneous breathing, the HFO ventilator was equipped with a custom-made DF system (DFS). The DFS is able to respond to fluctuations in mpaw on account of spontaneous breathing. The DFS is programmed to maintain a stable mpaw. During inspiration, fresh gas flow is increased, and during expiration, decreased. The DFS is capable of delivering fresh gas flow from 0 to 160 l/min. A more detailed description of the system is given in chapter 3.

Study protocol The study design is depicted in Figure 1. After 30 min of stabilization on conventional ventilation (CMV) after the last lung lavage, HFO ventilation was initiated. Initial settings: mpaw 20 cmH2O, proximal pressure amplitude (ΔP) was set to attain normocapnia (38–45 mmHg) and thereafter remained unchanged, oscillatory frequency 5 Hz, inspiration/expiration ratio 1:2, fresh gas flow 20 l/min and FiO2 1.0. HFO ventilation was then applied for 30 min in three different modes: 1, continuous fresh gas flow of 20 l/min and spontaneous breathing maintained (CF); 2, DF and spontaneous breathing maintained (DF), both in random order; and in a last step 3, during muscular paralysis (PAR). EIT and physiologic measurements were performed during the last 5 min of every HFO ventilation modality. To avoid the occurrence of lung collapse on account of the preceding HFO ventilation run, a recruitment maneuver was performed preceding each HFO ventilation modality.44,134 At the start of the recruitment maneuver, mpaw

Time random order

conventional ventilation high-frequency oscillatory ventilation

instrumentation lavage stabilization REC CF REC DFS REC paralyzed

Figure 1 Study design and electrical impedance tomography analysis. EIT, electrical impedance tomography measurement; REC, recruitment maneuver; ΔI, relative impedance change; Vt, tidal volume of spontaneous breathing.

Page 83: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 6

82

was increased to 30 cmH2O for five minutes. mpaw Was then reduced in steps of 3 cmH2O each 3 minutes until the animals started breathing in a regular pattern.

EIT measurements EIT measurements were performed using the Göttingen Goe-MF II EIT system (Cardinal Health, Yorba Linda, CA). Sixteen electrodes (Blue Sensor BR-50-K, Ambu, Denmark) were circumferentially applied around the chest at the xyphoid level. A 30 s reference measurement was recorded before lung lavage, Figure 1. All further measurements were referenced to this measurement. Measurements were performed at a scan rate of 44 Hz for 2 minutes. A 5mA peak-to-peak, 50 kHz electrical current was injected at one adjacent electrode pair and the resultant potential differences were measured at the remaining adjacent electrode pairs.

Figure 2 Electrical impedance tomography analysis. (A) Example of an electrical impedance tomography (EIT) image. Gray areas, within the white line, represent those pixels where the impedance variation exceeded 20% of the minimum pixel variation in the image, corresponding with the lung. (B) EIT image showing the tree regions of interest (ROIs) used for the evaluation of end expiratory lung volume (EELV). The total lung, total grey area. The ventral and dorsal lung regions, the light and dark grey areas, respectively. (C) Example of determination of fractional regional lung ventilation. Each half of the scan is analyzed in 32 ROIs. The 64 values represent the ventral to dorsal profiles of local ventilation in the chest as a percentage of the total ventilation of each half of the EIT image. The center of ventilation was defined as the point where the sum of fractional ventilation was 50% of the total fractional ventilation.

Page 84: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Regional lung characteristics during HFO ventilation: an animal study

83

Subsequently, all adjacent electrode pairs were used for current injection, thus completing one data cycle. A back-projection algorithm calculated the changes in impedance in time and reconstructed topographic EIT images of 912 pixels representing local impedance changes in a circular plane.147

EIT data analysis Both the respiratory and the cardiac components of the EIT signal were identified in the frequency spectra generated from all EIT measurements (Fourier transformation). To eliminate the cardiac signal from the impedance measurements, the cutoff frequency of the low-pass filter was set below the heart rate, 0.67 Hz, 40 beats/min.148 Figure 2A depicts the method used for assessing the region of interest of the lungs before lung lavage.149 For comparison of changes in end expiratory lung volumes (EELV), the changes in impedance ΔI were calibrated to Vt during CMV after the last lung lavage, Figure 1. For comparison of EELV, the change in EELV at the end of each HFO ventilation modality was referenced to the EELV at completion of the preceding recruitment maneuver. The EIT data were analyzed over three regions of interest, comprising the total area of the lungs and the upper and lower halves of the lungs, Figure 2B.150 The center of ventilation was determined in order to compare shifts in the distribution of ventilation between the three HFO ventilation modes. The center of ventilation was the point where the sum of fractional ventilation was 50% of the summed fractional ventilation, Figure 2C.151 To assess the occurrence of regional hyperinflation or recruitment on account of spontaneous breathing, regional filling characteristics of the lungs were calculated as depicted in Figure 3.152,153 To do this, the local relative impedance change was compared with the total relative impedance change of all pixels during inspiration and then fitted to a polynomial function of the second degree: y = ax2 + bx + c. See Figure 3 for a detailed explanation.

Hemodynamic and respiratory variables A MATLAB environment was used for data processing (MATLAB 7.04, The MathWorks, Natick, USA). The data processing is described in chapter 5. In short, in order to eliminate HFO ventilator oscillations, the recorded flow and pressure signals were low-pass filtered using a seventh-order Butterworth filter with a cutoff frequency of 2.5 Hz. The filtered flow signal represented the flow change caused by spontaneous breathing of the pigs. From the integrated filtered flow signals, breathing pattern and minute ventilation were determined for each individual

Page 85: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 6

84

Figure 3 Determination of regional filling characteristics. (A) Examples of relative impedance change in two individual pixels compared with the relative impedance change of all selected pixels representing the lungs. Both regional and global relative impedance change were calculated as a fraction of 1. Filling characteristics were calculated for every single spontaneous breath beginning at inspiration and ending at end-inspiration and averaged over the 2-min electrical impedance tomography (EIT) recording. • and represent the measured data points of one single spontaneous inspiration in two different pixels. The plots were fitted to a polynomial of the second degree, y = ax2 + bx + c, the black and grey lines. The polynomial coefficient of the second degree, a, describes the curve linearity of the plot. A polynomial coefficient a near zero (-0.2 to 0.2), near the dotted line, suggests a homogeneous regional tidal volume change on account of spontaneous breathing. A negative polynomial coefficient a (< 0.2) indicates low late regional filling and suggests local hyperinflation. A positive a (> 0.2) indicates low initial filling and suggests local recruitment. (B) Example of regional filling characteristics in the transverse plane in one animal. Polynomial coefficients of each individual pixel were averaged in each horizontal level in order to evaluate regional filling in the dorsal to ventral direction.153

Page 86: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Regional lung characteristics during HFO ventilation: an animal study

85

breath and averaged over a two-minute period. Arterial and mixed venous blood samples were analyzed using ABL505 and OSM3 hemoximeters (Radiometer, Copenhagen, Denmark). Continuous arterial blood gas analysis was conducted by the Paratrend 7 (Biomedical Sensors, High Wycombe, UK). Respiratory indices were calculated according to standard formulas.133

Statistical analysis Data are expressed as median and 25th to 75th interquartile range (IQR) unless otherwise stated. Parameter comparison for different HFO ventilation modes was performed using repeated measures analysis of variance with Bonferroni’s post hoc testing. Parameter comparison for the polynomial coefficient was performed using the Wilcoxon signed rank test. In all analyses, a p < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS 15 for Windows (SPSS, Chicago, IL).

Table 1 Respiratory and physiologic variables during different high-frequency oscillatory ventilation modes

spontaneously breathing

continuous flow demand flow paralyzed

Respiratory rate, (/min) 8.5 (7.8 - 9.4) 7.6 (6.4 - 9.9) a

Tidal volume, (ml/kg) 4.3 (4.2 - 4.7) 5.1 (4.2 - 6.4) a

Minute ventilation, (l/min) 2.1 (1.9 - 2.2) 2.2 (1.9 - 2.4)

mpaw, (cmH2O) 10 (9.9 - 11) 10 (8.7 - 11) 10 (9.9 - 11)

PaO2 start, (mmHg) 460 (454 - 498) 458 (421 – 509) 590 (536 - 611) a

PaO2 end, (mmHg) 493 (455 - 502) 473 (420-502) 451 (420 - 489)

ΔPaO2 , (mmHg) 19.9 (19.6 - 37) 31 (-5.5 to 84) -140 (-225 to -53) a

PaCO2 start, (mmHg) 54 (47 - 58) 55 (51 - 56) 56 (48 - 62)

PaCO2 end, (mmHg) 53 (52-54) 49 (46-52) 70 (63-72) a

ΔPaCO2, (mmHg) 6.3 (-3.0 - 18) -5.7 (-4.3 to -7.4) a 9.9 (3.6-22) a,b

a p < 0.05, vs. continuous flow; b p < 0.05 vs. demand flow. Data are expressed as median (IQR). mpaw, mean airway pressure; PaO2 start and PaO2 end, PaO2 at the start and at the end of the HFO ventilation modality; PaCO2 start and PaCO2 end, PaCO2 at the start and at the end of the HFO ventilation modality; ΔPaO2 and ΔPaCO2, difference in PaO2 and PaCO2 at the start and at the end of the HFO ventilation modality.

Page 87: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 6

86

Results

All animals completed the entire study protocol. Respiratory variables and data on gas exchange are summarized in Table 1. Vt of spontaneous breathing during HFO ventilation with DF was significantly higher, 5.1 ml/kg, compared with the Vt with continuous fresh gas flow, 4.3 ml/kg. Oxygenation improved in all HFO ventilation modes when spontaneous breathing was maintained. When pigs were paralyzed, PaO2 decreased remarkably during the 30-min experimental period. PaCO2 decreased using DF. PaCO2 increased significantly using a continuous fresh gas flow and during muscular paralysis.

Figure 4 Change in the end expiratory lung volume (EELV) determined from electrical impedance tomography (EIT) measurements for all high-frequency oscillatory (HFO) ventilation modalities. Data are expressed as median (IQR), *p <0.05. The EELV at the end of each HFO ventilation modality is referenced to the EELV at the start (EELVstart) of the same 30-min experimental run. A comparison of change in EELV was performed for the total lung EIT measurement and for the ventral and dorsal lung zones separately. CF, continuous fresh gas flow; DF, demand flow; PAR, paralyzed.

Figure 5 Ventral to dorsal distribution of lung ventilation. Data are expressed as median (IQR), * p < .05. Ventral to dorsal distribution of lung ventilation determined from EIT measurements for all HFO ventilation modalities. Center of ventilation is depicted for the left and right lung separately. CF, continuous fresh gas flow; DF, demand flow; PAR, paralyzed.

Page 88: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Regional lung characteristics during HFO ventilation: an animal study

87

Figure 4 depicts the changes in EELV referenced to EELV after the completion of each preceding recruitment maneuver. In all HFO ventilation modalities, a decrease in lung volume was observed. Lung volume was significantly better preserved when spontaneous breathing was maintained. When ventral and dorsal lung regions were considered separately, changes in lung volume were most markedly in the dorsal lung areas. In Figure 5, the center of lung ventilation, determined from functional EIT measurements for all HFO ventilation modalities, is shown for the right and the left lung separately. When the animals were breathing spontaneously on HFO ventilation and DF was used, the center of ventilation significantly shifted towards the dependent dorsal parts of both lungs compared with the measurements during muscular paralysis. Figure 6 summarizes the regional filling characteristics of all animals for both CF and DF. Summarizing all individual results, the polynomial coefficient calculated using HFO ventilation and continuous flow was -0.02 (IQR -0.05 to 0.14, range −0.35 to 0.32). For measurements using DF, polynomial coefficients were significantly different -0.01 (IQR -0.006 to 0.11, range -0.17 to 0.23), p = 0.01.

Discussion

The present study is the first to investigate the effect of spontaneous breathing during HFO ventilation on lung aeration and the distribution of ventilation. The main results of the animal experiment were that (1) lung volume was best preserved when spontaneous breathing was maintained during HFO ventilation,

Figure 6 Polynomial coefficients of regional versus global filling characteristics. Data are expressed as median (range), * p < .05. Polynomial coefficients of regional versus global filling, on account of spontaneous breathing, in the dorsal to ventral direction. Results determined from EIT measurements during HFO ventilation with continuous flow (CF) or demand flow (DF).

Page 89: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 6

88

predominantly in the dorsal dependent lung zones. (2) The use of DF during HFO ventilation shifted the center of ventilation to the dependent lung zones when compared with HFO ventilation during muscular paralysis. (3) There was no indication that spontaneous breaths during HFO ventilation resulted in regional hyperinflation. When lung aeration was considered, a positive effect of spontaneous breathing during HFO ventilation was observed. Lung volume at the end of expiration of spontaneous breathing during HFO ventilation, both for CF and DF, was significantly higher compared with the measurements during muscular paralysis. The maintenance of EELV was most markedly in the dependent dorsal lung areas. An explanation for the results on lung aeration is that ventilation and aeration are distributed differently when comparing spontaneous breathing and controlled mechanical ventilation.66,67 The most important factor responsible is the diaphragm. The diaphragm consists of an anterior tendon plate and a posterior muscle section. An active diaphragm lowers pleural pressure. As a result, transpulmonary pressure increases, enabling better aeration of dependent lung regions close to the diaphragm. In addition, during spontaneous breathing, the active dorsal muscle section of the diaphragm shifts the ventilation to the dorsal lung areas. A positive effect on lung aeration was already observed in several animal and clinical studies, when during mechanical ventilation, spontaneous breathing contributed only 10–30% to the total minute ventilation.66-69 After lung recruitment, a decrease in the lung volume was observed in all HFO ventilation modalities, Figure 4. The decrease in the lung volume can be explained by the low mpaw that could be applied. Higher mpaw would have prevented lung collapse. However, high mpaw is known to induce apnea in pigs, representing a limitation of the study (personal communication H. Wrigge from the department of Anesthesiology and Intensive Care Medicine of the University of Bonn). The decrease in the lung volume only coincided with a decrease in PaO2 when pigs were paralyzed. When spontaneous breathing was maintained, the decrease in the lung volume did not lead to a decrease in PaO2. Considering the mechanisms of gas exchange during HFO ventilation spontaneous breathing is not necessary.64 The results indicate, however, that a lower mpaw may be required to achieve adequate gas exchange when spontaneous breathing is maintained. We observed a significant shift in the center of ventilation to the dependent lung zones in favor of HFO ventilation with DF compared with HFO ventilation during muscular paralysis. This observation is in agreement with other studies that investigated the effect of lung recruitment on lung mechanics and the distribution

Page 90: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Regional lung characteristics during HFO ventilation: an animal study

89

of ventilation.151 Thus, in the HFO ventilation with DF, the spontaneous breathing not only resulted in a better preservation of EELV, it also directed ventilation towards the dependent lung zones. The redirection of ventilation favored gas exchange, indicated by the increase in PaO2 and decrease in PaCO2 during HFO ventilation with DF. One of the basic principles of a lung-protective HFO strategy is the application of small tidal volumes, usually below the anatomical dead space.64 Spontaneous breathing during HFO ventilation caused considerably larger volumes in our animal model. To assess the occurrence of regional hyperinflation and recruitment, regional filling characteristics of the lungs on account of spontaneous breathing were analyzed, Figure 6. Regional filling characteristics of the lungs were more homogeneous when DF was used instead of continuous flow. The more homogeneous distribution can be derived from the fact that polynomial coefficients were closer to 0 throughout the lungs when the DF was used. The cutoff values for the polynomial coefficients to indicate hyperinflation, recruitment or homogeneous filling are arbitrary. In a study describing the regional filling characteristics in mechanically ventilated adults with acute respiratory failure, PaO2/FiO2 <300mmHg, a much broader heterogeneity of regional filling was found.153 In that study, minimal regional polynomial coefficients varied from −2.8 to −0.56 (median −1.16), and maximum coefficients varied from 0.58 to 3.65 (median 1.41). In comparison with these data, the regional filling characteristics we observed were more homogeneous. Despite the fact that the tidal volumes of spontaneous breathing were higher using DF, there was no indication that this caused hyperinflation. It is an important observation with respect to the earlier discussed shift of the center of ventilation towards the dorsal lung regions when DF was used. This shift is explained by enhanced ventilation of the dorsal lung regions, rather than hyperinflation of the anterior lung parts during the use of DF. The interaction between spontaneous breathing and mechanical ventilation differs in various ventilation modes. Modes like airway pressure release ventilation (APRV) or biphasic positive pressure (BIPAP) ventilation allow unrestricted spontaneous breathing. In assist-control ventilation, only some breathing effort is necessary to trigger the ventilation and assist a breath. The presented modification of the HFO ventilator with a DFS allows unrestricted spontaneous breathing. The optimal mode for delivering partial ventilatory support is much discussed.154 A recent prospective observational study, however, did not demonstrate

Page 91: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 6

90

a difference in outcome when APRV/BIPAP and assist/control ventilation were compared.155 Not all research is in favor of the preservation of spontaneous breathing during mechanical ventilation. Some studies suggest that the use of neuromuscular blocking agents in the early phase of ALI/ARDS may improve oxygenation and reduce inflammation. Gentle spontaneous breathing may be beneficial; it can be argued that vigorous spontaneous respirations are contraindicated for the injured lung. Forceful respiration efforts can impose stress on the lungs and aggravate VILI.72,156,157 During our experiments with only mild lung injury, we observed calm spontaneous respirations with a limited tidal volume of spontaneous breaths. The effects of the HFO ventilator with DF on spontaneous breathing pattern and effort in severe lung injury need further study. The animal study has limitations. As the application of high mpaw prevented the animals from breathing spontaneously, only mild lung injury could be induced. Therefore, only two single lung lavages with no specific target for lung injury could be performed. With the limited lung lavage experimental conditions are not completely stable.127 The PaO2 at the start of HFO ventilation during paralysis indicate that lungs improved during the experiment. Whether similar results would be observed during spontaneous in severe human ARDS, using an open lung strategy with higher mpaw, will require further research.

Conclusions

This animal study demonstrates that spontaneous breathing during HFO ventilation preserves lung volume and when DF was used improves ventilation of the dependent lung areas. No significant hyperinflation occurred on account of spontaneous breathing. These results underline the importance of maintaining spontaneous breathing during HFO ventilation and support efforts to optimize HFO ventilators to facilitate patients’ spontaneous breathing.

Page 92: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

Reflections on pediatric high-frequency oscillatory ventilation from a physiologic perspective

Martin C.K. Kneyber Marc van Heerde

Dick G. Markhorst

Respiratory Care in press

Page 93: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 7

92

Abstract

Mechanical ventilation using low tidal volumes has become universally accepted to prevent ventilator-induced lung injury. High-frequency oscillatory (HFO) ventilation allows pulmonary gas exchange using very small tidal volumes (1 − 2 ml/kg) with concomitant decreased risk of atelectrauma. However, its use in pediatric critical care varies between only 3%–30% of all ventilated children. This might be explained by the fact that the beneficial effect of HFO ventilation on patient outcome has not been ascertained. Alternatively, in contrast with present recommendations and use it can be questioned if HFO ventilation has been employed in its most optimal fashion. This includes the indications for and timing of HFO ventilation as well as determining the best oscillator settings. The first was addressed in one small randomized study showing that early use of HFO ventilation instead of rescue use was associated with improved survival. From a physiologic perspective the oscillator settings could be refined. Lung volume is the main determinant of oxygenation in diffuse alveolar disease, suggesting using an open-lung strategy by recruitment maneuvers although this is in practice not custom. Using such an approach, the patient can be oscillated on the deflation limb of the pressure−volume curve allowing less pressure required to maintain a certain amount of lung volume. Gas exchange is amongst others determined by the frequency and the oscillatory power setting controlling the magnitude of the membrane displacement. Experimental work as well as preliminary human data has shown that it is possible to achieve the smallest tidal volume with concomitant adequate gas exchange when oscillating at high frequency and high fixed power setting. Future studies are needed to validate these novel approaches and evaluate their effect on patient outcome. 

Page 94: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Reflections on pediatric HFO ventilation from a physiologic perspective

93

Introduction

Mechanical ventilation (MV) is intimately linked with the daily care of critically ill children admitted to the pediatric intensive care unit (PICU). Indications for MV include diffuse alveolar disease (DAD) including acute lung injury (ALI) or acute respiratory distress syndrome (ARDS). Although life-saving, MV is also linked with ventilator-induced lung injury (VILI) and the development of multiple system organ failure (MSOF).6 This has led to the concept of lung-protective ventilation (LPV) that has become standard of care nowadays.158 High-frequency oscillatory (HFO) ventilation is, at least theoretically, an ideal tool for LPV as it allows pulmonary gas exchange using very small tidal volume Vt and decreases the risk of atelectrauma.50,138,159-169 Animal studies have pointed out that HFO ventilation might be preferable over conventional mechanical ventilation (CMV) given its more beneficial effects on oxygenation, lung compliance, attenuation of the pulmonary inflammation and histologic injury, and better alveolar stability.170,171 HFO ventilation allows the decoupling of oxygenation and ventilation. Simplified, oxygenation is dependent on lung volume which is controlled by the continuous distending pressure (CDP). The CDP is depicted by the oscillator as mean airway pressure mpaw. CO2 clearance (V& CO2) is relatively independent of lung volume but influenced by oscillatory frequency f and the square of Vt , V& CO2 = f ∙ Vt

2.172-175 The SensorMedics 3100A/B HFO ventilator (Yorba Linda, CA, USA) is the most commonly used HFO ventilation device in pediatrics. With this system, pressure oscillations with a f of 3–15 Hz are superimposed upon a CDP in a square-wave manner. The CDP is generated by a fixed fresh gas flow/bias flow leaving the ventilator circuit by an expiratory balloon valve. A membrane superimposes high-frequency pressure oscillations around the CDP. The oscillatory pressure amplitude is highly attenuated over the endotracheal tube and the airways and results in the delivery of a very small Vt’s, usually lower than anatomical dead space.64 Because of this small Vt there is a decreased risk of entering the so-called non-safe zones within the pressure volume loop of the diseased lung.42 The The use of HFO ventilation in pediatrics critical care varies between 3% and 30% of all ventilated children.12,176-179 This relatively low use can be explained by several factors. First, lack of equipment or disbelief of the attending physician because of the absence of sound evidence of effect. Second and perhaps even more importantly, many aspects of pediatric HFO ventilation remain to be explored including amongst others the identification of patients that are most likely to benefit from HFO ventilation, timing of HFO ventilation (early versus rescue),

Page 95: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 7

94

Table 1 Summary of clinical experiences with high-frequency oscillatory ventilation in critically ill children

Ref Study period N Inclusion criteria Initial HFOV settings

Randomized clinical trials

180 3½ y 58 OI > 13 or pulmonary barotrauma

> grade 1

f: 5 – 10 Hz

ΔP: chest wall wiggle

181 1 m 16 ARDS f: weight-dependent

ΔP: 10 > Ppeak CMV

Cohort studies

103 6 y 53 Patients with DAD and SAD f: weight-dependent

ΔP: chest wall wiggle

182 3 y 21 Patients with ARDS with OI > 10 and

PaO2/FiO2 < 200

Unknown

183 5½ y 100 Not reported Unknown

184 45 m 312 Severe ARF (PaO2/FiO2 < 150) with PEEP ≥

8 cmH2O, and/or PaCO2 ≥ 60 mmHg

f: 8 – 10 Hz

ΔP: 40 cmH2O

185 10 y 13 Confirmed RSV bronchiolitis f: 8 – 12 Hz

ΔP: chest wiggle

186 1½ y 232 Not reported f: 5 – 10 Hz

ΔP: chest wall wiggle

187 5 y 66 Not reported f: weight-dependent

ΔP: chest wall wiggle

188 3 m 6 OI > 13 F weight-dependent

Power 40

189 4 y 20 Weight ≤ 35 kg, FiO2 > 0.6 f: weight-dependent

ΔP: chest wall wiggle

190 4 y 35 DAD and SAD f: weight-dependent

ΔP: chest wall wiggle

191 1½ y 19 Patients with ARDS with PaO2/FiO2 < 200 Not reported

192 Unknown 123 OI > 13, gross air leak, weigh < 35 kg f: weight-dependent

ΔP: chest wall wiggle

193 Unknown 26 ARDS, stratification by duration of

conventional ventilation

f: weight-dependent

ΔP: chest wall wiggle

1Authors reported a decrease in mortality over time; 220 patients were managed with HFOV, the remaining with high-frequency jet ventilation; 37 patients with managed with HFOV, the remaining with high-frequency jet ventilation; 4Overall survival is shown. Authors reported differences in survival rate depending upon the underlying cause of the acute respiratory failure.

Page 96: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Reflections on pediatric HFO ventilation from a physiologic perspective

95

RM Survival (%) Outcome predictor(s)

Randomized clinical trials

No 66 HFOV

vs. 59

OI at 24 hrs

No 71 HFOV

vs. 44

sICAM-1

Cohort studies

Yes 64 Not reported

Unk 52.4 Survival predicted by OI at 24 hrs

Unk 451 Not reported

No 74 Death predicted by pre-HFV OI ≥ 20 and failure to decrease by 20% at 6 hrs of HFV

Yes 100 Not reported

Yes 53.44 Death independently predicted by immunodefiency and OI at 24 hrs of HFOV. Chronic

lung disease independently predicted by presence of sepsis and OI at 24 hrs of HFOV

No 39.4 Presence of non-pulmonary organ failure associated with death

Yes 40 Not reported

Yes 75 Not reported

Yes 88.6 Not reported

Unk 73.7 Initial OI > 20 and failure to decrease by 20% at 6 hrs predicted death

No 41.7 In non-survivors OI increased after 24 hrs of HFV

Yes 42 Early HFOV (≤ 24 hrs) associated with significant improvement in mortality

OI, oxygenation index; HFOV, high-frequency oscillatory ventilation; y, year; m, month; rm, recruitment maneuver; f, frequency; ΔP, proximal pressure amplitude; CMV, conventional mechanical ventilation; ARDS, acute respiratory distress syndrome; DAD, diffuse alveolar disease; SAD, small airway disease; ARF, acute renal failure; unk, unknown.

Page 97: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 7

96

optimal oscillator settings, and monitoring during HFO ventilation. The purpose of this perspective is to review published clinical experiences, animal and bench studies with HFO ventilation and to reflect on how its use might be improved in light of the physiological properties of the diseased lung.

Clinical experiences

The effect of HFO ventilation on mortality was compared with conventional MV (CMV) in two randomized controlled trials (RCTs), Table 1.180,181 The largest of the two was performed years ago in five centers during a 3½ year period.180 In this cross-over study 58 patients with acute respiratory failure or pulmonary barotrauma, and an oxygenation index (OI) > 13 demonstrated by two consecutive measurements over a 6 hour period were randomized to either HFO ventilation (n = 29) using a strategy of aggressive increase in CDP targeted at a SpO2 ≥ 90% with a FiO2 ≤ 0.6, or CMV (n = 29) using a strategy utilizing positive end expiratory pressure (PEEP) and limited inspiratory pressures. Patients with obstructive airway disease, intractable septic or cardiogenic shock or non-pulmonary terminal diagnosis were excluded. Targeted blood gas values were equal for each group. The main finding was that HFO ventilation did not improve survival (HFO ventilation 66% vs. 59%) or total ventilator days (HFO ventilation 20 ± 27 vs. 22 ± 17) compared with CMV when the data were analyzed by initial assignment. However, the percentage of survivors requiring supplemental oxygen at 30 days was significantly lower in the HFO ventilation group (21% vs. 59%, p = 0.039). Furthermore, mortality was only 6% (n = 1/17) in patients who were exclusively managed on HFO ventilation, whereas it was 42% (n = 8/19) for patients who failed CMV and were transitioned to HFO ventilation. Yet, mortality in patients who were exclusively managed with CMV was 40% (n = 4/10). Samransamruajkit reported the results of a small single-center study comparing HFO ventilation (n = 7 patients) with CMV (n = 9 patients) with ARDS in a two-year study period.181 Survival was higher with HFO ventilation (71%) compared with CMV (44%) and predicted by plasma levels of soluble intercellular adhesion molecule (sICAM) 1. Both RCTs have not been repeated so far, but various institutions have described their (limited) experiences with HFO ventilation, Table 1.103,182-191,194 Overall survival varied between 40% and 90%. The largest cohort study came from a collaborative of 10 pediatric centers reporting 232 patients.186 Duration of CMV

Page 98: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Reflections on pediatric HFO ventilation from a physiologic perspective

97

prior to HFO ventilation was between 2.2 ± 4.2 to 4.5 ± 3.1 days whereas patients with pre-existing lung injury were managed for up to 11.4 ± 45.5 days before transfer to HFO ventilation. 30-day mortality ranged from 30% for patients with respiratory syncytial virus (RSV) lower respiratory track disease (LRTD) to 59% for patients with congenital heart disease. Mortality was independently predicted by the OI 24 hours after start of HFO ventilation and the presence of immunocompromise. The applicability of the OI as a predictor for patient outcome during HFO ventilation has been confirmed by others.182,192 Some have linked failure of the OI to improve by at least 20% six hours after transition to HFO ventilation with adverse outcome.184,191 The use of HFO ventilation in pulmonary conditions with increased airway resistance and prolonged time constants, such as virus-induced obstructive airway disease (OAD), remains subject of debate because of the assumed risk of dynamic air-trapping as seen in HFJV. Nevertheless, several institutions have reported safe and beneficial use of HFO ventilation in this patient population.103,185,186,188,190,195 It can thus be concluded that at present a beneficial effect of HFO ventilation on mortality has not been established. This may be explained by various factors. First, the knowledge on LPV has significantly increased over the past years. It is now universally accepted that a low Vt should be applied. However, the study by Arnold and colleagues was conducted in the era prior to the ARDS Network trial. In their study, the authors did not specify the Vt’s used on CMV. Similar criticisms can be made towards the study by Samransamruajkit so that it is not unthinkable that patients on CMV were subjected to high Vt. Second, both RCTs were not powered to detect statistical significant differences in mortality.

Critical appraisal of the HFO ventilation strategy employed

Alternatively, the question could also be raised whether HFO ventilation was applied in its most optimal fashion. These issues (amongst others) include identification of the patient who will benefit the most from HFO ventilation, the timing of cross-over from CMV to HFO ventilation, as well as determining the best oscillator settings and weaning from HFO ventilation. These will subsequently be discussed.

Indications for and timing of HFO ventilation The indications for HFO ventilation are ill-defined and usually depend upon the personal preference of the attending physician. In general, HFO ventilation is only

Page 99: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 7

98

considered as a rescue approach when CMV fails. One group of investigators have evaluated the early use of HFO ventilation instead of using it as rescue therapy.193 In their small observational study of 26 patients it was found that the group of patients that was transitioned to HFO ventilation within 24 hours of CMV had a significantly higher 30-day survival rate (58.8 vs. 12.5%). We suggest that HFO ventilation should be considered if oxygenation remains severely impaired (in our institutions defined by a SpO2 < 88% and/or a PaO2 < 50 mmHg with FiO2 > 0.6) despite the application of maximal lung-protective CMV, i.e. limiting inspiratory pressures to 30−35 cmH2O and a sufficient level of PEEP, in children with ALI/ARDS. Alternatively, the OI can be used although a specific threshold needs to be determined. For patients with OAD there is no guideline available when to consider HFO ventilation. Based upon our own experiences we consider HFO ventilation for these patients when refractory respiratory acidosis persists despite maximum conservative measures such as nebulization or intravenous administration of bronchodilators, use of heliox or use of external PEEP to stent occluded airways. In our opinion there are no known contra-indications for HFO ventilation. Although its safety has been questioned in patients with severe traumatic brain injury (sTBI) based upon the assumption that the high intrathoracic pressures are propagated towards the brain and impede the cerebral circulation. However, this has been refuted by both animal and clinical data.196,197

Best HFO ventilation approach and oscillator settings for oxygenation Lung volume is the main determinant of oxygenation in DAD during HFO ventilation. The PaO2 increases linearly with lung volume.198 This suggests that it is imperative to employ an open lung strategy (i.e. opening up the lung and keeping it open) in DAD by (repeated) recruitment maneuvers (RM). Animal work has indeed shown improved lung compliance and less hyaline membrane formation when such strategies were applied.163,199,200 Furthermore, pressure oscillations are less dampened in lungs with ongoing atelectasis, thus exposing the conducting airways to higher injurious pressure swings.201 However, in both pediatric RCTs as well as in nearly half of all observational cohort studies there is no mention of RMs being performed.180-184,187,191,192 Also, there is much ongoing scientific debate related to RMs. Not all lung diseases are recruitable and in general the potential for lung recruitability is highly variable.202 Furthermore, there are so far no clinical studies establishing the beneficial effects of RMs during HFO ventilation let alone that it is fully understood what the best RM is. The latter has been addressed in one study in

Page 100: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Reflections on pediatric HFO ventilation from a physiologic perspective

99

which four different RM approaches were compared: a step-wise pressure increase over 6 mins, a 20 s sustained dynamic inflation either one or repeated six times, and a standard approach (setting mpaw direct at start).203 This study showed that a step-wise pressure increase produced the greatest increase in lung volume and resolution of atelectasis. However, it could not be ascertained if re-expansion of the lung was complete albeit that final lung volume with this approach was close to that of the deflation limb of the pressure–volume (P–V) curve. Thus, this study suggests that the stepwise increase pressure approach might be considered for optimizing lung volume during HFO ventilation as it incorporates not only pressure but also adequate duration of the RM. The clinical benefits of RMs during HFO ventilation have been addressed in a recently completed phase II trial in critically ill adults comparing HFO ventilation with and without RMs (www.clinicaltrials.gov, NCT00399581). Unfortunately, a pediatric counterpart is lacking but the adult results are eagerly awaited. Another, at least theoretical, benefit of RMs is that it allows oscillating the patient on the deflation limb of the P–V curve thereby (partially) avoiding injurious hyperinflation and atelectasis.42,45,65,134,204-207 By doing so, less CDP is needed to maintain a certain lung volume on the deflation limb compared with the inflation limb because of the hysteresis of the respiratory system. In our view, this can be achieved in clinical practice in patients with DAD by initially setting the CDP 3−5 cmH2O above the mpaw on CMV as the distal CDP is lower than the set proximal CDP.208,209 Then the CDP should be increased stepwise over a certain period of time until the point where oxygenation (either the SpO2 or the PaO2) does not improve or the proximal pressure amplitude (ΔP) starts to increase (this point is assumed to represent total lung capacity).210 Subsequently, the CDP is reduced to the point where oxygenation and ΔP worsens again after initial improvement. A positive effect of sustained inflations prior to the stepwise increase in CDP has not been ascertained.138,203 It is of importance to understand that RMs in patients with OAD are different from patients with DAD. For patients with OAD, the purpose of the stepwise increase in CDP is to splint open and stent the airways. From a clinical perspective, this point is presumed when the PaCO2 starts to drop. Then the CDP is not increased any further as relatively healthy alveoli face the risk of being exposed to a high mpaw once the airways are open.210   

Page 101: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 7

100

Best HFO ventilation approach and oscillator settings for ventilation The V& CO2 is determined by patient-related characteristics and oscillator settings. The first include compliance (Crs) and resistance of the respiratory system (Rrs).211,212 With reduced compliance, in unresolved atelectasis, there is a marked increase in transmission of the peak-to-through ΔP to the alveoli and bronchi. Increased Rrs decreases the transmission of the peak-to-through ΔP over the airways to the alveoli.211 Oscillator settings include oscillatory power setting (magnitude of membrane displacement), frequency f in Hertz (Hz), inspiratory to expiratory ratio (I:E ratio), position of the membrane, endotracheal tube (ETT) length and diameter, and presence of ETT leakage.175,213,214 The endotracheal tube constitutes the major work load to the oscillator and is an important determinant of Vt.215,216 Vt is proportional to the ETT inner cross-sectional area because the impedance of the ETT exceeds the impedance of the lung.217,218 Increasing diameter (ID 2.5 to 4.0 mm) of the ETT increases pressure transmission.211 The manufacturer’s manual recommends setting f and power according to the patients’ age, ventilator settings and observation of chest wiggle. This recommendation has been adopted into clinical practice, using the f and power in a weight and age-dependent manner in both RCTs as well as the observational cohort studies.103,180,181,184-190,192,193 We propose that these recommendations can be refined. From a physiological perspective it seems rational to use the highest possible f in DAD. First, f determines the rate of oscillations and directly influences the Vt. Hence, the higher the f the smaller the Vt because changes in f are inversely proportional to the distal oscillatory pressure amplitude. Consequently, it requires less effort to stay within the limits of the safe zone (i.e. the zone with the smallest risk of injurious hyperinflation or atelectasis) of the P–V loop . Second, collapsed lung regions are more easily opened at higher f.219 Third, the delivered Vt is more equally distributed as it becomes less dependent on regional compliance at higher f.220 Lastly, the square block wave form is better preserved allowing a more constant Vt.51,221 The next question then is what could be considered as an optimal f. Venegas et.al. have proposed that how the f needs to be set depends upon the so-called corner frequency fc of the lung, fc = 1/(2πRC), where R is resistance and C compliance.65 fc defines the optimal frequency at which there is adequate gas transport during HFO ventilation in combination with the least injurious pressures, and is influenced by the underlying disease, Figure 1. The fc is increased in lung diseases characterized

Page 102: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Reflections on pediatric HFO ventilation from a physiologic perspective

101

by short time constants and low compliance such as in DAD. This implies that at higher f alveoli are ventilated at a lower pressure cost of ventilation as opposed to lung diseases characterized by prolonged time constants (for example obstructive airway disease). Importantly, f is intimately linked with ΔP. Basically, the higher the ΔP the larger the Vt. Yet, we (unpublished data) and others have observed in bench test studies that Vt was smaller when combining high f (15 Hz) and high power (set to achieve a ΔP of 90 cmH2O) compared with low f (5 Hz) and low power settings as the distal pressure amplitude was much lower.222 Importantly, V& CO2 was still sufficient with the combination of high f and high power. These findings were in agreement with the work from Hager and co-workers. They have measured Vt in adult patients with ARDS managed on HFO ventilation and found smaller Vt with the combination high f and high power setting.216 The use of these higher f did not impair gas exchange.223 Importantly, how are these theoretical benefits translated into clinical practice? At present it is impossible to detect the fc and thus impossible to identify the optimal f. Furthermore, what ΔP should be targeted for? Based upon our own experiences we propose using the highest f in combination with a fixed power setting that is associated with acceptable CO2 elimination (in our view pH > 7.25) in patients with DAD. For patients with OAD the f should be chosen much lower (for instance 5−7 Hz). It has been argued that the ΔP should not exceed 70–90 because higher pressures may theoretically expose the proximal airways to injurious pressures. Theoretically, the use of high amplitudes might lead to gas trapping due to the development of so-called choke points causing expiratory flow limitation,

Figure 1 Corner frequency for different underlying lung conditions. Corner frequency fc of the lung in patients with decreased compliance such as ALI/ARDS, increased resistance such as obstructive airway disease and normal lungs. fc, graphically depicted by the dot, defines the optimal frequency at which there is adequate gas transport during HFO ventilation in combination with the least injurious pressures. It is defined by 1/2πRC, where R is resistance and C compliance. Adapted from Venegas and Fredberg.51,65

Page 103: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 7

102

especially at low CDP.224 However, the occurrence of choke points has never been demonstrated.

Monitoring during HFO ventilation At present physicians only have the SpO2, blood gas analysis, ΔP and chest radiography at their disposal for evaluating the response of a patient to HFO ventilation. Developments are being made with respect to electrical impedance tomography (EIT) and respiratory inductance plethysmography (RIP) incorporated in the Bicore II™ as tools for the determination of the optimal CDP.225,226 We have recently begun to explore the use of RIP in guiding the stepwise increase in CDP. Alternatively, the optimal CDP may be recognized when both lung compliance and oxygenation index (OI, calculated by mpaw ∙ FiO2 ∙ 100/PaO2) are optimal.227 The benefit of the OI over the PaO2/FiO2 ratio is that it takes the degree of ventilator settings, as summarized by the mpaw, into account. Van Genderingen and co-workers found that the lowest OI during the RM indicated at which CDP the oxygenation was considered to be optimal; this also indicated the point on the deflation limb of the P–V curve where physiologic shunt fraction was the lowest.228 The oscillatory pressure ratio (OPR) may also aid in the identification.212 OPR is defined as the ratio of the distal and proximal ETT pressure swings. To calculate the OPR it is necessary to measure the tracheal pressure. In a 3.0 mm ETT neonatal IRDS simulated model OPR decreased when the CDP was increased suggestive for lung recruitment, but increased when the CDP was increased further. This suggested hyperinflation. The OPR was the lowest at maximum compliance. The OPR was also affected by frequency, ΔP and ETT ID. The OPR was further evaluated in an animal model of acute lung injury.229 One of the main findings of this study was that, after lung recruitment, similar oxygenation with smaller pressure swings could be achieved with a lower CDP set by the deflation limb of the P–V curve rather than the inflation limb. The clinical use of these potential aids however needs to be established.

Weaning from HFO ventilation In our opinion it is possible to extubate a child without going back to CMV. However, this requires the child to breathe spontaneously. Maintaining spontaneous breathing during HFO ventilation improves oxygenation and regional ventilation.143,230 Spontaneous breathing during HFO ventilation is feasible for small children but becomes more difficult when the patient demands high inspiratory flows. The maximal possible bias flow delivered by the oscillator may be well below

Page 104: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Reflections on pediatric HFO ventilation from a physiologic perspective

103

the needs of the patient. This will lead to increased imposed work of breathing as shown by our group in a bench test model.112 Because of this many older patients on the oscillator are likely to need sedatives and neuromuscular blockade during their illness that prohibits spontaneous breathing. Nonetheless, current adult protocols attempt to maintain spontaneous breathing during HFO ventilation.113

Conclusions

The beneficial effect of HFO ventilation on outcome in critically ill children remains unclear. However, it can be questioned if HFO ventilation has been employed in its most optimal fashion. In contrast with present recommendations and use, we suggest for patients with diffuse alveolar disease to convert to HFO ventilation early in the disease course, employ an open-lung strategy using (repeated) recruitment maneuvers and oscillate at the highest possible frequency and high fixed power setting that is associated with adequate gas exchange. For patients with obstructive airway disease HFO ventilation can be used to open up and stent the airways. Future prospects include validation of these approaches.

Page 105: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging
Page 106: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

General Discussion

Page 107: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 8

106

The main objective of this thesis was to optimize the HFO ventilator design (SensorMedics, 3100 A/B, Yorba Linda, CA, USA) to better tolerate spontaneous breathing during HFO ventilation. Spontaneous breathing during HFO ventilation impedes ventilator function, and we hypothesized that the HFO ventilator imposes a high additional workload to a spontaneous breathing subject. In a bench test we gained insight into factors contributing to the imposed work of breathing (WOB) during HFO ventilation. We identified the fixed and rather low fresh gas flow as an important contributing factor. It led to the development of a demand flow system (DFS) for the SensorMedics HFO ventilator. The performance of the DFS was assessed in several bench tests. These tests were focused on the capability to reduce work of breathing, and evaluated the response time and pressurization of the DFS. The DFS was also evaluated in a series of animal experiments. First the effect of the DFS on WOB was studied. In addition the effects of spontaneous breathing during HFO ventilation on gas exchange and regional lung characteristics were evaluated. Based on what was known already and with new insights gained from recent studies on HFO ventilation we proposed a novel approach to pediatric HFO ventilation. This PhD project started in 2006. Since 2006 many research groups published studies on mechanical ventilation. An important topic was how to deal with spontaneous breathing during mechanical ventilation. Highlights on this topic are discussed here.

What was learned from the first bench test

The currently used SensorMedics HFO ventilators are able to provide gas exchange in an apneic patient. HFO ventilation, as part of lung protective ventilation strategies, focused on limiting delivered tidal volumes as much as possible. How to deal with spontaneous breathing during HFO ventilation was not a focus in designing the HFO ventilators. The difficulties of spontaneous breathing during HFO ventilation in adult patients were long recognized.231 The solution of this problem was first sought in suppression of breathing by deep sedation or even muscular paralysis. As the beneficial effects of spontaneous breathing during mechanical ventilation gained interest, we made efforts to optimize the current HFO ventilator design to better tolerate spontaneous breathing. First a bench test was performed to assess the imposed WOB of spontaneous breathing during HFO ventilation. The bench test evaluated the influence of patient and equipment related factors on the amount of imposed WOB: breathing patterns

Page 108: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General discussion

107

for different patient sizes, endotracheal tube size and the fresh gas flow rate. Especially the fresh gas flow rate has a clinical relevant influence on the amount of imposed WOB. A high fresh gas flow rate reduces the imposed WOB significantly. In order to find the optimal approach to tolerate spontaneous breathing during HFO ventilation, the HFO ventilator can be considered as a modified continuous positive airway pressure (CPAP) system. In a HFO ventilator pressure oscillations are applied on a constant mean airway pressure mpaw or CPAP. CPAP systems can be classified in several groups: continuous flow and demand flow systems. A HFO ventilator can for clarification be considered as a continuous flow CPAP system. In a HFO ventilator the mpaw or CPAP is generated by a continuous fresh gas flow and the expiratory balloon valve. WOB is extensively studied in different types of CPAP systems.125,129,232-234 Continuous high flow CPAP systems, with an elastic reservoir in the inspiratory limb, need a fresh gas flow rate of 50 to 100 l/min for imposed WOB relief. An advantage of continuous flow CPAP systems is that the patient does not need to trigger valves. In a demand flow CPAP system the patient needs to trigger valves. As a result the imposed work of breathing can be higher in demand flow CPAP systems compared to continuous flow systems. More recent generations of ventilators have minimized imposed effort by more responsive demand systems and more aggressive flow delivery systems.122 Based on the findings of the first bench test we advised to set the fresh gas flow rate during HFO ventilation to the maximum possible flow rate of 60 l/min. Some clinicians experimented with the use of high gas flow rates during HFO ventilation in a clinical setting. In one center the HFO ventilators were modified to be able to increase the fresh gas flow rate even above 60 l/min, data not published. It had however a detrimental effect on gas exchange in some patients. Some caution is therefore needed when the maximum continuous fresh gas flow rate is used. The effect of bias flow rate on gas exchange during HFO ventilation is not much studied. A certain mpaw in the HFO ventilator circuit can be set by different combinations of fresh gas flow rate and expiratory valve resistance. In order to maintain a mpaw at high fresh gas flow rates, the expiratory balloon valve resistance has to be set low. A low expiratory resistance reduces the impedance of the HFO ventilator circuit. The reduced circuit impedance attenuates the high-frequency pressure oscillations. As a result gas exchange on account of high-frequency oscillations may become ineffective. This can be an explanation for the worsened gas exchange seen when using high fresh gas flow rates.

Page 109: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 8

108

The remaining solution then, to better tolerate spontaneous breathing during HFO ventilation, was the development of a demand flow system.

Bird’s-eye view on the demand flow system control design

The demand flow system for the HFO ventilator had, from a clinical point of view, to fulfill several conditions. The system had to relief work of breathing and provide sufficient synchrony with a spontaneous breathing patient. Furthermore the demand flow system had to improve ventilator function during spontaneous breathing. The mathematical description of the demand flow systems control, chapter 3, may appear complex and intimidating to clinicians not familiar with control theory. Therefore a bird’s-eye view on the controller design is provided here. Also, from the ‘modern theory of control’ point of view, the control algorithm for the demand flow system had to fulfill several criteria. The algorithm should operate in real time with a short response time. High pressure overshoots need to be avoided because of patient safety. The algorithm need to be robust. That means that the properties of the controller do not change much in different conditions for different patients with different lung diseases. It also needs to be robust for all possible HFO ventilator settings. Furthermore control accuracy and stability are required for the control algorithm. The linear-quadratic-Gaussian (LQG) controller that was used is a combination of a Kalman filter with a linear-quadratic (LQ) regulator. These are solutions to what probably are the most fundamental problems in control theory. A LQG controller can provide an optimal solution for control of a dynamic system such as the HFO ventilator connected to a breathing patient.235 The Kalman filter is a mathematical method named after Rudolf E. Kalman.236 The Kalman filter was originally developed for use in spacecraft navigation and turned out to be useful for many applications: medical monitoring, financial econometrics, object tracking. A good filtering algorithm can remove noise (random variations) from signals while retaining the useful information. The Kalman filter is mainly used to estimate system states that can only be observed indirectly or inaccurately by the system itself, Figure 1. Its purpose is to use measurements observed over time, containing noise and other inaccuracies, and produce values that tend to be closer to the true values of the measurements and their associated calculated values.237

Page 110: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General discussion

109

The purpose of the demand flow system is to regulate the fresh gas flow rate in order to reject disturbances in mean airway pressure caused by spontaneous breathing. The airway pressure is the only measured output available to adjust the demand flow to a patients need. The airway pressure signal contains however measurement noise and other inaccuracies. These inaccuracies are for example caused by the high-frequency pressure oscillations. Regulation of the demand flow system using only the measured airway pressure alone is therefore not appropriate. Here the Kalman filter provides the solution for dealing with the noise and inaccuracies. The Kalman filter produces estimates of the true paw of both measurements of paw and their associated calculated values of paw. This is done by predicting a value of paw, estimating the uncertainty of the predicted paw, and computing a weighted average of the predicted paw and the measured paw. The most weight is given to the value of paw with the least uncertainty. The estimates of paw produced by the Kalman filter tend to be closer to the true paw than the original measurements because the weighted average has a better estimated uncertainty than either of the values that went into the weighted average, Figure 2. Simplified for the demand flow system is the current estimation of paw:

K ∙ Measured value of paw + (1 – K) ∙ Previous estimation of paw, where paw is the airway pressure and K the Kalman gain, Figure 2. The Kalman filter finds the optimum averaging factor, Kalman gain, for each consequent state of the, HFO ventilator/patient, system. The demand flow system measures paw at a sampling rate of 200/s. The current estimate is based on the actual measurement

Figure 1 Typical Kalman filter application.

Page 111: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 8

110

and the previous estimation 1/200s earlier. For each time step the Kalman gain changes. In order to use a Kalman filter to remove noise from the paw signal, the process that is measured must be described in a system model. For the demand flow system a model was created using an electro-acoustic analogy. In this model, a resistor–capacitor (RC) circuit, together with pressure and gas-flow sources, were used to approximate the behavior of the patients’ respiratory system, the HFO ventilator, and the ventilator circuit. To meet all clinical requirements the demand flow system has to eliminate fluctuations in paw caused by spontaneous breathing. The LQ regulator was used for this disturbance rejection. The LQ regulator adjust the demand flow based on input from the measured paw, Kalman filter and the actual demand flow rate. The functioning of the demand flow system was evaluated in several bench tests and animal experiments.

Demand flow system evaluation & effects of spontaneous breathing

The demand flow system performance was evaluated in the studies described in chapter 3, 4 and 5. The effects of spontaneous breathing during HFO ventilation were evaluated in the animal experiments described in chapter 5 and 6.

Figure 2 Graphical depiction of the Kalman filter principle. The current estimation of the airway pressure paw is a weighted average of the current measurement and the previous estimation. To calculate the current estimation, the uncertainty or standard deviation of the current measurement (σmeas) and previous estimation (σpe) are required. The Kalman filter provides these standard deviations. Note that the current estimation has a smaller standard deviation (σce) than either σmeas and σpe, which is to say that the uncertainty in the estimate of paw is decreased by combining the two pieces of information.

Page 112: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General discussion

111

Work of breathing Work is required to breathe. Respiratory system disease can increase the work required to realize sufficient gas exchange. Mechanical ventilation can reduce the increased effort. When spontaneous breathing is maintained during mechanical ventilation, the ventilator may however impose an additional workload to the patient. Conventional ventilators are designed to minimize the imposed workload and reduce the total work of breathing. A workload in the physiologic range is considered optimal.82 Full unloading, for instance reducing the total WOB to zero, induces loss of respiratory muscles. Excessive respiratory muscle loading may cause muscle fatigue and weaning failure. The bench test described in chapter 2 demonstrated that the HFO ventilator imposed a significant workload during spontaneous breathing in HFO ventilation simulations. The bench test in chapter 4, with the demand flow system, showed a 38–59% reduction in the imposed workload, as assessed by the pressure time product (PTP).  These results were promising since the amount of imposed work of breathing (WOB) was lowered to a level in the physiologic range. The results obtained from the bench test need to be interpreted with some caution. In their seminal paper ‘mechanics of breathing in man’, Otis, Fenn and Rahn in 1950 described how ventilation in man is regulated.135 Respiratory rate and tidal volume are adopted so that the respiratory effort is minimized. This implies that an imposed workload influences the spontaneous breathing pattern. In a bench test the interactions of a breathing subject with a ventilator are difficult to simulate. The results of the animal experiments provide therefore more valuable results in this respect. The demand flow system was able to totally reject the workload imposed by the HFO ventilator on the animals.

Demand flow system synchrony Effective relief of undesirable high respiratory muscle loading in ventilated patients suffering from ALI/ARDS is important. The importance of ventilator-patient synchrony should in this context not be underestimated.137,238 A ventilator can provide excellent relief of work of breathing during a whole respiratory cycle, dissynchrony may however be present. For example when initially much effort is needed to trigger the ventilator and after that the ventilator fully compensates for this effort during the rest of the respiratory cycle. Since the introduction of the SensorMedics HFO ventilators in the 1980s much effort was invested in patient-ventilator interaction of mechanical ventilators. The

Page 113: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 8

112

HFO ventilator equipped with a demand flow system had in our opinion in this respect to be comparable with modern ventilators. The demand flow system trigger performance is similar to the performance of conventional ventilators currently used in clinical practice. The triggering delay of the demand flow system was 50 ms and is as fast as modern conventional mechanical ventilators. The inspiratory delay was 115 ms and comparable with delay reported for conventional ventilators, chapter 3.122 Some caution is needed when interpreting these results. For offline analysis of the trigger performance it is necessary to filter the relevant pressure signals in order to remove high frequency oscillations. This mathematical filtering affects the filtered pressure signal. The applied filtering technique directly influences the results for the different response times. For that reason trigger performance was reported only for measurements without oscillations. During the animal experiments no clinical signs of dissynchrony were present. Animals were breathing in a calm regular pattern and no active expirations were observed.

HFO ventilator functioning In the standard configuration of the SensorMedics HFO ventilators spontaneous breathing causes large fluctuations in mpaw. These pressure swings interrupt oscillations, activates alarms and can even lead to ventilator shut off. The demand flow system fully eliminated these problems. By providing a stable mpaw, alarms were silenced and automatic ventilator shut off was avoided. During clinical use of the standard HFO ventilator it is known that mpaw can deviate from the set mpaw over time. This is mainly caused by fluctuations in expiratory balloon valve resistance. Because of open loop control of this valve, in the original HFO ventilator design, no automatic adjustments are made to correct deviations of the mpaw from the set mpaw. Therefore alarms are set to prevent the occurrence of an injurious high or low mpaw. The demand flow system compensates for changes in expiratory valve resistance by adjusting the fresh gas flow rate. As a result deviations from set mpaw over time are prevented.

Gas exchange & regional lung characteristics In the animal experiments the results for gas exchange show a benefit of spontaneous breathing during HFO ventilation. PaO2 improved in all HFO ventilation modes when spontaneous breathing was maintained. Muscular paralysis had a deteriorating effect on PaO2. The combination of spontaneous breathing and use of the demand flow system lowered the PaCO2. These findings are in accordance with the results obtained with electrical impedance tomography

Page 114: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General discussion

113

(EIT) in chapter 6. This study demonstrates that spontaneous breathing during HFO ventilation preserves lung volume and, when DF was used, improved ventilation of the dependent lung areas. When spontaneous breathing is maintained during HFO ventilation it is important to evaluate whether this coincides with (regional) over inflation. In the animal experiments no significant hyperinflation occurred on account of spontaneous breathing.

Paralysis in acute respiratory distress syndrome: back to the future?

Optimizing gas exchange, avoiding lung injury, and preserving respiratory muscle strength and endurance are vital therapeutic objectives for managing ALI and ARDS. Accordingly, comparing the physiology and consequences of breathing patterns during mechanical ventilation that preserve or eliminate breathing effort is a theme of persisting investigation. The recent introduction of ventilation techniques geared toward integrating natural breathing rhythms into even the earliest phase of ARDS support, e.g. airway pressure release ventilation (APRV) and neurally adjusted ventilatory assist (NAVA), has stimulated investigation.239 Research was focused on both spontaneously regulated but also controlled ventilation. Studies on the use of paralytic drugs, in the early phase of ALI/ARDS support, provided important insights. Especially vigorous spontaneous breathing can have adverse effects, muscular paralysis may prevent these effects. To balance the enthusiasm for the preservation of spontaneous breathing, some comments on the role of muscular paralysis need to be made. When patient triggered ventilation is applied in the early phase of ARDS, vigorous respiration can be seen. Forceful inspiratory and also expiratory respiratory muscle activity are elicited by a high minute ventilation. The high demand is caused by a combination of agitation, disease caused increase in oxygen consumption and an increased alveolar dead space. Besides, the use of PEEP may cause expiratory muscle contraction to oppose lung expansion. Especially active expiration may therefore cause loss of lung volume by cephalad movement of the diaphragm. Muscular paralysis eliminates inspiratory and expiratory respiratory muscle activity. In recruitable lungs paralysis may cause higher end expiratory lung volumes and improve oxygenation.240,241 A diseased lung can boost respiratory muscle oxygen demand. When fully controlled mechanical ventilation is applied, all energy is provided by the

Page 115: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 8

114

ventilator. In fully controlled ventilation besides muscular paralysis, there is a need for sedative and analgesic drugs. This will further reduce oxygen consumption.242,243 To preserve spontaneous breathing when demands are high, such as in the early phase of ARDS, may not be the preferred strategy.72,156 In a recent multi-center trial, Papazian showed that muscular paralysis in the first 48 hours of severe ARDS, PaO2/FiO2 < 150, was associated with decreased mortality.157 The crude 90-day mortality did not significantly differ. After adjustment for baseline characteristics, PaO2/FiO2, simplified acute physiology score (SAPS) II and plateau pressure, improved 90 day survival rate was seen when treated with the neuromuscular blocking agent. Controlled ventilation implies the need for heavy sedation and muscular paralysis. It has a negative impact on respiratory muscle strength and also on cognitive behavior. Only 18 hours complete diaphragmatic inactivity and mechanical ventilation results in marked atrophy of human diaphragmatic myofibers.244 Delirium is more frequently seen in deeply sedated patients.245,246 On the positive side, some evidence exists that nondepolarizing paralytic drugs have direct anti-inflammatory effects.72 The meaning of this in the occurrence of VILI and multi system organ failure remains unclear. Very potent anti-inflammatory drugs did not show any effect on mortality in the treatment of ALI-ARDS.247 Another intriguing and new concept in lung protective ventilation, is the airway propagation hypothesis and its consequences.248 Inflammatory lung injury may occasionally begin focally and propagate via the airways from distal to proximal. In the initial stages of lung injury, the first 24–48 hours, alveolar edema is highly mobile. Spreading of this mediator-laden and protein rich fluid can injure previously normal lung regions. Deep breathing, frequent position change and coughing are commonly seen in spontaneous regulated ventilation. These factors contribute to propagation of lung injury. To prevent propagation controlled ventilation may therefore, in theory, be preferred. Totally opposed to the use of muscular paralysis and heavy sedation is to use no sedation at all for critically ill patients receiving mechanical ventilation.249 Strøm showed in a single center randomized trial that no sedation, only analgesia, in ventilated patients was associated with an increase in days without ventilation and a significant reduction in ICU and hospital length of stay. Patients with an expected need of mechanical ventilation of more than 24 hours were included in this trial.

Page 116: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

General discussion

115

To breathe or not to breathe? That is the question

The trials conducted by Papazian and Strøm studied totally different patient populations, severe ARDS vs. patients requiring mechanical ventilation for more than 24 hours. The consequences of the results are much discussed. Put together they revitalize the question: ‘To breathe or not to breathe’ during mechanical ventilation. The patients with severe ARDS, as in the Papazian trial, we usually consider as the ideal candidates to benefit from HFO ventilation. The Papazian trial underlines the importance not to be dogmatic about preservation of spontaneous breathing. Muscular paralysis has its place in the early phase of ARDS, when demands are high. The use of demand flow during HFO ventilation however, may in our opinion decrease the necessity of muscular paralysis and/or heavy sedation. The Strøm trial compared the effect on two sedation protocols on weaning from mechanical ventilation. Preservation of spontaneous breathing during mechanical ventilation was not a focus of the study. The trial however emphasizes the role of sedation in weaning from mechanical ventilation. The intervention group received no sedation, only analgesia. The control group received intravenous sedation, propofol and midazolam, and underwent daily interruption of sedation to assess extubation readiness.250 In this light, when demand flow enables the preservation of spontaneous breathing during HFO ventilation, the need for sedation may be less and consequently weaning is facilitated. It seems reasonable to consider paralysis in the early phase of only severe ARDS and to preserve spontaneous breathing in all other cases. Since the start of this project in 2006 several clinical studies tried to find the optimal ventilator mode to incorporate spontaneous respiration in a mechanical ventilation strategy.155,251,252 No mode proved to be the most advantageous yet.

Perspectives

Since the introduction of HFO ventilation in the early 1980s there is a continuing debate of its place in the treatment of patients with respiratory failure. There is still controversy regarding how to best optimize HFO ventilation to achieve optimal gas exchange and lung-protection. Based on a strong physiological rationale, rapidly expanding use internationally, and promising results in early small RCTs, a definitive RCT, to establish the impact of HFO ventilation versus best current conventional ventilation on mortality, has started enrolment of adult patients in

Page 117: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 8

116

2009. We look forward to the results of this Oscillation for ARDS Treated Early (OSCILLATE) trial. What else remains to be done in the perspective of this thesis? Our studies show that it is possible to support spontaneous breathing during HFO ventilation. An important limitation was that we were not able to evaluate the demand flow system and the effects of spontaneous breathing during HFO ventilation in ALI/ARDS conditions. The effects of spontaneous breathing in lung conditions necessitating a high mpaw, such as in severe ARDS, need to be studied. Especially since the Papazian trial, using neuromuscular blockade in the early phase of severe ARDS, showed a reduction in adjusted mortality in the treatment group. The optimal way to wean patients from HFO ventilation is not much studied. HFO ventilation with demand flow may prove to be valuable here. In conventional mechanical ventilation controlled modes are mostly used in the initial phase of MV, followed by some kind of assisted mode in the weaning phase. Weaning from HFO ventilation completely on the HFO ventilator may be better possible when spontaneous breathing is supported by demand flow. The sometimes tricky switch, back to conventional ventilation in the weaning phase may be prevented in this way. Introduction of HFO ventilation with demand flow in the weaning phase seems a good first step into introduction in a clinical setting. Weaning from mechanical ventilation starts directly after intubation. The different aspects of weaning from mechanical ventilation are extensively studied in the adult patient population. A lot of progress is made in this area. More studies are definitely needed for the pediatric patient population specifically. Focus of research has to be aimed at ventilator patient synchrony and protocol-based weaning. Next to this, optimal sedation protocols, weaning from sedation during mechanical ventilation and, for example, daily interruption of sedation need to be studied for the pediatric population. HFO ventilation strategies have a sound theoretical basis. The effects of HFO ventilator settings on different parameters among others, delivered Vt, and regional lung characteristics are very well studied in a research setting. These parameters are however not available at the bedside in daily practice, but would make a valuable contribution in optimizing HFO ventilation. Electrical impedance tomography (EIT) can provide global and regional information on lung mechanics. EIT analysis is time consuming at the moment. Efforts need to be made to optimize EIT systems in order to obtain real time information on the lungs.

Page 118: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter

 

Summary

Page 119: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 9

118

Chapter 1

Mechanical ventilation is a mainstay in modern intensive care. It is used for the treatment of patients suffering from respiratory failure. In the nineteenth century already, doubts were raised about the safety of positive pressure ventilation. Since the 1950s extensive research showed that mechanical ventilation itself can induce or aggravate lung injury. This phenomenon is known as ventilator-induced lung injury (VILI). Animal and clinical studies on acute respiratory distress syndrome (ARDS) and acute lung injury (ALI) have elucidated the different mechanisms responsible for VILI. The term barotrauma was introduced after the observation that high airway pressures caused air leaks in the lungs. Further research demonstrated that high airway pressures are not inevitably injurious to the lungs. High airway pressures are injurious when they cause high tidal volumes, named volutrauma. In fact barotrauma and volutrauma are linked by the pressure volume relation of the lungs. Atelectrauma describes the lung injury that is induced by repetitive opening and closing of distal lung units during the respiratory cycle. Besides mechanical forces also the inflammatory response produces VILI, this is called biotrauma. Inflammatory cells and mediators play an important role in the development of lung injury. Overspill of these mediators in the systemic circulation can lead to multiple system organ failure. Once all these mechanisms responsible for VILI were elucidated, ventilator strategies were sought that could protect the already injured lung from additional harm. High-frequency oscillatory (HFO) ventilation is one of these lung protective ventilation strategies. HFO ventilation delivers pressure oscillations, with a frequency between 3–15 Hz, around a constant mean airway pressure, producing small tidal volumes of around 1–2 ml/kg body weight. These tidal volumes are often less than the anatomical dead space. During HFO ventilation, the tidal volumes and associated swings in alveolar pressure are very small. When applied in a recruited lung, this approach should theoretically limit ventilator induced lung injury. The preservation of spontaneous breathing during artificial ventilation gained interest in lung protective ventilation strategies. Maintenance of spontaneous breathing in mechanically ventilated patients has several beneficial effects. Spontaneous breathing augments ventilation perfusion matching and cardiopulmonary function, reduces sedative requirement and shortens intensive care stay. Current lung protective ventilation protocols therefore aim at maintenance of spontaneous breathing. Preservation of spontaneous breathing

Page 120: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Summary

119

during mechanical ventilation was not yet an issue during the development of the HFO ventilator (SensorMedics, 3100 A/B, Yorba Linda, CA, USA) in the 1970s and 1980s. In HFO ventilation, more conventional respiratory rates and tidal volumes, as in spontaneous breathing efforts, are not needed to achieve adequate gas exchange. Unfortunately, with current HFO ventilator design vigorous respiratory efforts in large pediatric and adult patients may cause pressure swings that activate alarms, interrupt oscillations, and produce significant desaturations. Initial HFO ventilation trials in adults recommended muscular paralysis for this reason. The main objective of this thesis was to optimize the HFO ventilator design to better tolerate spontaneous breathing during HFO ventilation.

Chapter 2

When during mechanical ventilation spontaneous breathing is maintained, the mechanical ventilator and ventilator circuit impose workload to the patient. In a bench test, we investigated which factors contributed to the imposed work of breathing (WOB) in a SensorMedics 3100 A/B HFO ventilator. A computer-controlled piston-driven test lung was used to simulate a spontaneously breathing patient. The test lung was connected to a HFO ventilator by an endotracheal tube. The spontaneous breath rate and volume, tube size and ventilator settings were simulated as representative of the newborn to adult range. The fresh gas flow rate was set at a low and a high level. The imposed WOB was calculated using the Campbell diagram. The main result of this study is that the imposed WOB can be markedly increased during HFO ventilation in pediatric and adult patients, especially at low fresh gas flow rates. This can be a good explanation for the discomfort seen in patients breathing spontaneously during HFO ventilation. The fresh gas flow rate and peak inspiratory flow are both strongly related to the imposed WOB. In the pediatric and adult simulations, the imposed WOB exceeded the normal physiologic WOB by as much as 400%. Besides, the mean airway pressure was not maintained in the breathing circuit when the inspiratory flow exceeded the fresh gas flow rate, this even led to ventilator shutdown.

Page 121: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 9

120

Chapter 3

The low and fixed fresh gas flow rate is an important factor in the HFO ventilator design responsible for the high workload imposed to a spontaneously breathing patient. Fluctuations in mean airway pressure also impede ventilator functioning. To overcome this problem a demand flow system was developed for the HFO ventilator. Aim of the demand flow system is to compensate for changes in mean airway pressure on account of spontaneous breathing. The basic principle of the demand flow system is to increase the fresh gas flow during inspiration and to decrease flow during expiration. As a result the imposed WOB decreases. In chapter 3 a description is provided of the hardware part and the control algorithm of the demand flow system. The hardware part of the demand flow system consists of an electronically controlled mass flow valve, proximal pressure measurement sensor with a necessary electric circuit and control and communication electronics. The control software is developed in a MATLAB/Simulink environment (The MathWorks, Natick, USA). Using a linear quadratic Gaussian (LQG) state feedback controller, the demand flow system alters the fresh gas flow in the ventilator circuit in response to spontaneous breathing.

Chapter 4

The demand flow system was first evaluated in a bench test. Again a computer-controlled piston-driven test lung was used to simulate a spontaneously breathing patient. Spontaneous breathing was simulated to represent shallow and normal to deep breathing for a large pediatric or adult patient at a normal and rapid breath rate. The Campbell diagram and pressure time product (PTP) were used to quantify the imposed work of breathing. Using the demand flow system, imposed WOB is considerably reduced. The demand flow system reduces inspiratory imposed WOB by 30% to 56% and inspiratory imposed PTP by 38% to 59% compared to continuous fresh gas flow. Expiratory imposed WOB was decreased as well by 12% to 49%. In simulations of shallow to normal breathing for an adult, imposed WOB is 0.5 J/l at maximum. A WOB level in the physiological range, approximately 0.5 J/l in adults, seems to correspond with an optimal workload for the respiratory muscles. Fluctuations in mean airway pressure on account of spontaneous breathing are markedly reduced when the demand flow system was used. Furthermore simulation of vigorous spontaneous breathing did not trigger

Page 122: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Summary

121

ventilator alarms and did not lead to ventilator shutdown, when using demand flow.

Chapter 5

The breathing pattern of a subject during mechanical ventilation is highly dependent on the interaction with and characteristics of the ventilator. For that reason the evaluation of spontaneous breathing during mechanical ventilation in a bench test has important limitations. Patient ventilator interaction cannot adequately be evaluated in a bench test. Therefore the HFO ventilator with demand flow system was evaluated in an animal model of mild lung injury. In eight pigs (47-64 kg) lung injury was induced by lung lavage with normal saline. After spontaneous breathing was restored HFO ventilation was applied, in runs of 30 minutes, with continuous fresh gas flow (CF) or the demand flow system operated in two different setups. Pressure to regulate the demand flow system was sampled directly at the Y-piece of the ventilator circuit (DFS) or between the endotracheal tube and measurement equipment at the proximal end of the endotracheal tube (DFSPROX). In the end, animals were studied paralyzed. Breathing pattern, work of breathing, and gas exchange were evaluated. HFO ventilation with demand flow decreased breathing frequency and increased tidal volume compared with CF. Comparing HFO modes CF, DFS, and DFSPROX, total pressure time product (PTP) was 66 cmH2O/s/min (interquartile range 59−74), 64 cmH2O/s/min (50−72), and 51 cmH2O/s/min (41−63). Ventilator PTP was 36 cmH2O/s/min (32−42), 8.6 cmH2O/s/min (7.4−10), and 1 cmH2O/s/min (−1.0 to 2.8). Oxygenation, evaluated by PaO2, was preserved when spontaneous breathing was maintained and deteriorated when pigs were paralyzed. Ventilation, evaluated by PaCO2, improved with demand flow. PaCO2 increased when using continuous flow and during muscular paralysis. We concluded that in moderately lung-injured anesthetized pigs during HFO ventilation, demand flow facilitated spontaneous breathing and augmented gas exchange. Demand flow decreased total breathing effort as quantified by PTP. Imposed work caused by the HFO ventilator appeared totally reduced by demand flow.

Page 123: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Chapter 9

122

Chapter 6

In the same series of experiments described in chapter 5, we evaluated the effect of spontaneous breathing on regional lung characteristics during HFO with the use of electrical impedance tomography (EIT). EIT is a noninvasive technique for pulmonary imaging. EIT is able to accurately describe both global and regional lung volume changes over time during mechanical ventilation. Compared to electron beam computed tomography, as an established method to assess changes in local air content, simultaneous EIT measurements correlate quite closely. EIT was used to assess: regional lung aeration and ventilation and the occurrence of hyperinflation on account of spontaneous breathing during HFO ventilation. End expiratory lung volume (EELV) was best preserved when spontaneous breathing was maintained during HFO ventilation compared to HFO ventilation during muscular paralysis. Lung volume was predominantly preserved in the dependent lung regions. A significant shift in ventilation toward the dependent lung regions was observed in spontaneously breathing animals on HFO ventilation with DF compared to HFO ventilation and spontaneous breathing suppressed. No signs of regional hyperinflation on account of spontaneous breathing was observed with either CF or DF.

Chapter 7

Chapter 7 provides a novel approach of high-frequency oscillatory ventilation in the pediatric population. The approach is a consensus of three based on current insights in the application of high-frequency ventilation in the pediatric population.

Chapter 8

This chapter provides a general discussion. The selected approach for the development of the demand flow system is explained. A simplified description of the control algorithm for the demand flow system is provided. The major findings of the studies presented in this thesis are discussed. The concept of spontaneous breathing during mechanical ventilation is discussed based on new insights gained from recent clinical studies. Future perspectives are given.

Page 124: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Hoofdstuk

 

Samenvatting

Page 125: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Hoofdstuk 9

124

Hoofdstuk 1

Beademing is één van de pijlers van de moderne intensive care. Beademing wordt toegepast in de behandeling van patiënten met respiratoir falen. Al in de 19e eeuw bestond er twijfel over de veiligheid van positieve druk beademing. Sinds de jaren vijftig van de 20e eeuw is uit uitvoerig onderzoek gebleken dat kunstmatige beademing op zichzelf long schade kan veroorzaken of verergeren. Dit fenomeen wordt beademing geïnduceerde longschade genoemd. Dierexperimenteel en klinisch onderzoek hebben de mechanismen verantwoordelijk voor het ontstaan van beademing geïnduceerde longschade verhelderd. De term barotrauma werd geïntroduceerd na de observatie dat door het opleggen van een hoge luchtweg druk luchtlekkage ontstond in longblaasjes. Nader onderzoek toonde aan dat een hoge luchtweg druk niet onvermijdelijk resulteert in schade. Een hoge luchtweg druk is schadelijk als dit leidt tot een hoog long volume. Dit wordt met volutrauma aangeduid. In feite zijn de begrippen baro- en volutrauma aan elkaar gekoppeld door de druk volume relatie van de long. Het begrip atelectase trauma beschrijft de longschade die ontstaat door herhaaldelijk openen en samenvallen van longblaasjes tijdens ademhalingscycli. Naast de genoemde mechanische factoren zijn ontstekingsreacties verantwoordelijk voor het ontstaan van beademing geïnduceerde longschade, dit wordt aangeduid met biotrauma. Ontstekingscellen en ontstekingsmediatoren spelen een belangrijke rol in het ontstaan van longschade. Deze ontstekingsreactie blijft niet alleen beperkt tot de long. Via de circulatie bereiken deze ontstekingsmediatoren andere organen. Deze mediatoren induceren zo schade aan deze organen en kunnen uiteindelijk in multi orgaan falen resulteren. Een toegenomen inzicht in het ontstaan van beademing geïnduceerde longschade heeft geresulteerd in beademingsstrategieën die de long beschermen tegen additionele schade als gevolg van de beademing. Hoogfrequente oscillerende (HFO) beademing is een van deze beademingsstrategieën. Tijdens HFO beademing wordt beademd met hoogfrequente druk oscillaties rondom een constante gemiddelde luchtweg druk. De constante luchtwegdruk zorgt voor het openen en open houden van longblaasjes in de zieke long. De druk oscillaties liggen in het frequentiebereik van 3 tot 15 Hz, en resulteren in een teugvolume van rond de 1 tot 2 ml/kg lichaamsgewicht. Dit teugvolume is lager dan het volume van de anatomisch dode ruimte. De kleine teugvolumes gaan schade door overmatige rek van de long tegen. In theorie is HFO beademing de volmaakte long protectieve beademingstherapie.

Page 126: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Samenvatting

125

Het behoud van spontane ademhaling tijdens beademing heeft diverse gunstige effecten. Behoud van de spontane ademhaling geeft een meer optimale verdeling van ventilatie en doorbloeding van de long. Daarnaast verbetert het de functie van het hart, vermindert de behoefte aan sedatie en wordt de duur van beademen en ook opname verkort. De huidige strategieën voor long beschermende beademing hebben dan ook als doel de spontane ademhaling van een patiënt te behouden. Ten tijde van de ontwikkeling van de huidige HFO beademingsapparatuur (SensorMedics, 3100 A/B, Yorba Linda, Californië, VS) in de jaren 70 en 80 van de 20e eeuw werd nog geen rekening gehouden met het behoud van spontane ademhaling. De manier waarop gaswisseling tijdens HFO beademing tot stand komt verschilt essentieel van gaswisseling tijdens conventionele beademing. Daardoor zijn meer conventionele ademfrequenties en ademteugen, zoals tijdens spontaan ademen, niet noodzakelijk voor een adequate gaswisseling tijdens HFO beademing. Krachtige spontane ademhaling van een groot kind of volwassene geeft met het huidige ontwerp van de HFO beademingsapparatuur zelfs problemen. Het belemmert het functioneren van de HFO beademingsapparatuur. Een krachtige ademhaling veroorzaakt grote schommelingen in de gemiddelde luchtwegdruk, dit activeert alarmen, onderbreekt de oscillaties, resulteert soms zelfs in desatureren van de patiënt en uitschakelen van het apparaat. De eerste klinische studies met HFO beademing adviseerden daarom spontane ademhaling te onderdrukken door middel van diepe sedatie of zelfs spierverslapping. Het voornaamste doel van de studies beschreven in dit proefschrift is het ontwerp van de HFO beademingsapparatuur zo te verbeteren dat spontane ademhaling van een patiënt goed wordt verdragen.

Hoofdstuk 2

Indien tijdens beademing de spontane ademhaling behouden blijft, dan veroorzaken de beademingsmachine, beademingscircuit en de endotracheale tube extra ademarbeid naast de normale fysiologische ademarbeid. In een proefopstelling met een SensorMedics 3100 A/B HFO beademingsapparaat hebben we geëvalueerd welke factoren in welke mate bijdragen aan deze additionele ademarbeid. Het belangrijkste resultaat van deze studie is dat de additionele ademarbeid zeer hoog is indien spontaan ademen tijdens HFO beademing van het grote kind of een volwassen patiënt word gesimuleerd. En dan met name indien er een lage toevoer

Page 127: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Hoofdstuk 9

126

van beademingsgas is. Dit is een goede verklaring voor het discomfort dat gezien wordt bij patiënten die tijdens HFO beademing spontaan ademen. De continue toevoer van beademingsgas en de piek inspiratoire luchtstroom waren de belangrijkste determinanten van de additionele ademarbeid. In simulaties voor grote kinderen en volwassenen overtrof de additionele ademarbeid de normale fysiologische ademarbeid met meer dan 400%. Als gevolg van het spontane ademen werden fluctuaties in de gemiddelde luchtwegdruk gezien. Indien de piek inspiratoire luchtstroom hoger was dan de toevoer van beademingsgas werd hierdoor het HFO beademingsapparaat zelfs uitgeschakeld.

Hoofdstuk 3

In het ontwerp van de HFO beademingsapparaat bleek de lage en niet variabele toevoer van beademingsgas een belangrijke factor verantwoordelijk voor de hoge additionele ademarbeid. Een mogelijke oplossing voor dit probleem is gebruik te maken van een variabele toevoer van beademingsgas. De variatie in toevoer van beademingsgas dient daarbij afgestemd te zijn op de behoefte van de patiënt. Doel van een systeem met een variabele toevoer van beademingsgas is te compenseren voor door spontane ademhaling veroorzaakte schommelingen in de gemiddelde luchtwegdruk. Tijdens inademen van een patiënt wordt de toevoer van beademingsgas in het HFO circuit verhoogd, tijdens uitademen verlaagd. In hoofdstuk 3 wordt een beschrijving gegeven van de apparatuur en het controle algoritme van het systeem voor variabele gastoevoer tijdens HFO beademing. De apparatuur bestaat uit een elektronische regelklep, druksensor voor het meten van de luchtwegdruk, elektronische circuits en controle en communicatie elektronica. Het controle algoritme is ontwikkeld in een MATLAB/Simulink software omgeving (The MathWorks, Natick, VS). Met gebruik van een lineaire kwadratische Gaussische (LQR) regelaar wordt de toevoer van beademingsgas tijdens HFO beademing afgestemd op de spontane ademhaling.

Hoofdstuk 4

Het systeem voor variabele toevoer van beademingsgas tijdens HFO beademing werd allereerst getest in een proefopstelling. Gelijk aan de testopstelling zoals beschreven in hoofdstuk 2 werd spontaan ademen gesimuleerd met behulp van een computer gestuurde zuigerpomp. In deze opstelling werd oppervlakkig, normaal en

Page 128: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Samenvatting

127

diep ademen van een groot kind en een volwassene gesimuleerd. Het Campbell diagram en het druk tijd product werden gebruikt om de additionele ademarbeid te evalueren. Het systeem voor variabele toevoer van beademingsgas verminderde de additionele ademarbeid tijdens de inspiratie met 30–56% volgens het Campbell diagram en met 38–59% berekend met het druk tijd product. De expiratoire additionele ademarbeid werd met 12–49% gereduceerd. In simulaties van een oppervlakkige rustige ademhaling van een volwassene was de maximale additionele ademarbeid 0.5 J/l. Een additionele ademarbeid gelijk aan de normale fysiologische ademarbeid, ongeveer 0.5 J/l voor volwassenen, lijkt overeen te komen met een optimale belasting van de ademhalingsspieren. Het volledig wegnemen van alle ademarbeid, leidt tot verlies van ademhalingsspieren. Overmatige belasting van de ademhalingsspieren leidt tot uitputting van deze spieren en zelfs het mislukken van het ontwennen van de beademing. De gemiddelde luchtwegdruk was stabieler tijdens simulatie van spontane ademhaling bij gebruik van het systeem met een variabele beademingsgas toevoer. Bij simulaties van krachtige spontane ademhaling werden daardoor geen alarmen geactiveerd en resulteerde dit niet in het uitschakelen van de HFO apparatuur.

Hoofdstuk 5

Het ademhalingspatroon van een patiënt tijdens kunstmatige beademing is zeer afhankelijk van de interactie met en de eigenschappen van het beademingsapparaat. Interactie van een patiënt met een beademingsapparaat kan daardoor onvoldoende geëvalueerd worden in een proefopstelling. Dit was aanleiding om de HFO beademing met een variabele toevoer van beademingsgas te testen in een diermodel met geringe longschade. Bij acht varkens (47−64kg) werd milde longschade geïnduceerd door herhaalde long lavage met fysiologisch zout. Bij de spontaan ademende dieren werd HFO beademing gestart, telkens gedurende 30 minuten, met een constante beademingsgas toevoer (CF) en met variabele beademingsgas toevoer. De druk voor het aansturen van het systeem van variabele gas toevoer werd op twee verschillende punten in het beademingscircuit gemeten. Deze druk werd of ter hoogte van het Y stuk van het beademingscircuit gemeten (DFS) of proximaal tussen de endotracheale tube en de meet apparatuur (DFSPROX). Aan het einde van elk experiment werden metingen verricht bij verslapte dieren. We toonden aan dat in dit diermodel, met milde longschade, HFO beademing met een variabele gas toevoer de spontane ademhaling ondersteunt en tevens leidt tot

Page 129: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Hoofdstuk 9

128

een verbetering in gaswisseling. Een variabele gas toevoer verlaagt de totale ademarbeid. De additionele ademarbeid veroorzaakt door de HFO beademings-apparatuur kan door de variabele gas toevoer geheel worden weggenomen.

Hoofdstuk 6

In hetzelfde experiment, zoals beschreven in hoofdstuk 5, evalueerden we het effect van spontane ademhaling tijdens HFO beademing op regionale longdelen. Hierbij maken we gebruik van elektrische impedantie tomografie (EIT). EIT is een niet invasieve techniek voor het afbeelden dynamische processen in de longen. Te denken valt aan veranderingen in luchthoudendheid en ventilatie van de long. Met EIT kunnen zowel globale als regionale veranderingen in de long geëvalueerd worden. De CT-scan heeft zich bewezen als een goede techniek om bijvoorbeeld de luchthoudendheid van de long te beoordelen. Gelijktijdig uitgevoerd EIT onderzoek komt goed met deze CT-scans overeen. De studie in hoofdstuk 6 beschrijft met behulp van EIT: veranderingen in regionale luchthoudendheid en ventilatie van de long als gevolg van spontane ademhaling tijdens HFO beademing. Daarnaast is gekeken of de spontane ademhaling tijdens HFO beademing leidt tot schadelijke volume toename van delen van de long. Het eind expiratoire longvolume wordt het best gehandhaafd indien de spontane ademhaling behouden blijft tijdens HFO beademing. Met name aan de rugzijde van de longen. Door spierverslapping neemt het eind expiratoire longvolume het meest af. Bij behoudt van spontane ademhaling en het gebruik van de variabele beademingsgas toevoer verschuift de ventilatie van de long naar de rugzijde vergeleken met spierverslapte dieren. Ademhalen tijdens HFO beademing resulteerde niet in schadelijk uitrekken van de long, in tegendeel. Het teugvolume werd meer homogeen in de long verdeeld.

Hoofdstuk 7

Dit hoofdstuk voorziet in een praktische aanbeveling voor de klinische toepassing van HFO beademing op de kinderleeftijd. Het protocol is een consensus van drie gebaseerd op de huidige inzichten in HFO beademing in deze leeftijdscategorie.

Page 130: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Samenvatting

129

Hoofdstuk 8

In dit hoofdstuk worden de voornaamste bevindingen bediscussieerd. De technische benadering van het probleem van spontane ademhaling tijdens HFO beademing wordt uiteengezet. Het controle algoritme van de variabele beademingsgas toevoer wordt vereenvoudigd weergegeven. De belangrijkst conclusies van de diverse studies worden weergegeven. Het concept van behoud van spontane ademhaling tijdens beademing wordt bediscussieerd in het licht van nieuwe inzichten verkregen uit recent onderzoek. Tot slot wordt de richting voor toekomstig onderzoek uiteengezet.

Page 131: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging
Page 132: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

Page 133: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

132

References

1 Jayne WA (1925). The Healing Gods Of Ancient Civilizations. New Haven: Yale University Press.

2 Vasalius A (1543). De vivorum sectione nonnulla. In: De humani corporis fabrica, pp. 662. Basel, Johannes Oporinus.

3 Leroy J (1828). Second mémoire sur l'asphyxie. J Physiol Exp Pathol, 8, 97-135.

4 The Edinburgh Almanack of Universal Scots and Imperial Register (1834). Edinburgh: Oliver & Boyd.

5 Lassen HC. (1953). A preliminary report on the 1952 epidemic of poliomyelitis in Copenhagen with special reference to the treatment of acute respiratory insufficiency. Lancet, 1, 37-41.

6 Tremblay LN, Slutsky AS (2006). Ventilator-induced lung injury: from the bench to the bedside. Intensive Care Med, 32, 24-33.

7 Ashbaugh DG, Bigelow DB, Petty TL, Levine BE (1967). Acute respiratory distress in adults. Lancet, 2, 319-323.

8 Brun-Buisson C, Minelli C, Bertolini G, Brazzi L, Pimentel J, Lewandowski K et al. (2004). Epidemiology and outcome of acute lung injury in European intensive care units. Results from the ALIVE study. Intensive Care Med, 30, 51-61.

9 Rubenfeld GD, Caldwell E, Peabody E, Weaver J, Martin DP, Neff M et al. (2005). Incidence and outcomes of acute lung injury. N Engl J Med, 353, 1685-1693.

10 Vincent JL, Sakr Y, Ranieri VM (2003). Epidemiology and outcome of acute respiratory failure in intensive care unit patients. Crit Care Med, 31, S296-S299.

11 Trachsel D, McCrindle BW, Nakagawa S, Bohn D (2005). Oxygenation index predicts outcome in children with acute hypoxemic respiratory failure. Am J Respir Crit Care Med, 172, 206-211.

12 Flori HR, Glidden DV, Rutherford GW, Matthay MA (2005). Pediatric acute lung injury: prospective evaluation of risk factors associated with mortality. Am J Respir Crit Care Med, 171, 995-1001.

13 West JB (2003). Thoughts on the pulmonary blood-gas barrier. Am J Physiol Lung Cell Mol Physiol, 285, L501-L513.

14 Brigham KL, Woolverton WC, Blake LH, Staub NC (1974). Increased sheep lung vascular permeability caused by pseudomonas bacteremia. J Clin Invest, 54, 792-804.

15 Ohkuda K, Nakahara K, Weidner WJ, Binder A, Staub NC (1978). Lung fluid exchange after uneven pulmonary artery obstruction in sheep. Circ Res, 43, 152-161.

16 Brigham KL, Woolverton WC, Staub NC (1974). Reversible increase in pulmonary vascular permeability after Pseudomonas aeruginosa bacteremia in unanesthetized sheep. Chest, 65, Suppl-54S.

17 Fein A, Grossman RF, Jones JG, Overland E, Pitts L, Murray JF et al. (1979). The value of edema fluid protein measurement in patients with pulmonary edema. Am J Med, 67, 32-38.

18 Greene KE, Wright JR, Steinberg KP, Ruzinski JT, Caldwell E, Wong WB et al. (1999). Serial changes in surfactant-associated proteins in lung and serum before and after onset of ARDS. Am J Respir Crit Care Med, 160, 1843-1850.

19 Ware LB, Matthay MA (2000). The acute respiratory distress syndrome. N Engl J Med, 342, 1334-1349.

20 Donald KW (1947). Oxygen poisoning in man; signs and symptoms of oxygen poisoning. Br Med J, 1, 712-717.

21 Donald KW (1947). Oxygen poisoning in man. Br Med J, 1, 667.

22 Macklin CC (1939). Transport of air along sheaths of pulmonic blood vessels from alveoli to mediastinum: clinical implications. Archives of Internal Medicine, 64, 913-926.

23 Greenfield LJ, Ebert PA, Benson DW (1964). Effect of positive pressure ventilation on surface tension properties of lung extracts. Anesthesiology, 25, 312-316.

Page 134: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

133

24 Fowler WS (1949). Lung function studies; uneven pulmonary ventilation in normal subjects and in patients with pulmonary disease. J Appl Physiol, 2, 283-299.

25 Mead J, Takishima T, Leith D (1970). Stress distribution in lungs: a model of pulmonary elasticity. J Appl Physiol, 28, 596-608.

26 Webb HH, Tierney DF (1974). Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures. Protection by positive end-expiratory pressure. Am Rev Respir Dis, 110, 556-565.

27 Dreyfuss D, Soler P, Basset G, Saumon G (1988). High inflation pressure pulmonary edema. Respective effects of high airway pressure, high tidal volume, and positive end-expiratory pressure. Am Rev Respir Dis, 137, 1159-1164.

28 Dos Santos CC, Slutsky AS (2000). Invited review: mechanisms of ventilator-induced lung injury: a perspective. J Appl Physiol, 89, 1645-1655.

29 Dreyfuss D, Saumon G (1998). Ventilator-induced lung injury: lessons from experimental studies. Am J Respir Crit Care Med, 157, 294-323.

30 Muscedere JG, Mullen JB, Gan K, Slutsky AS (1994). Tidal ventilation at low airway pressures can augment lung injury. Am J Respir Crit Care Med, 149, 1327-1334.

31 Hickling KG (2002). Reinterpreting the pressure-volume curve in patients with acute respiratory distress syndrome. Curr Opin Crit Care, 8, 32-38.

32 Hamilton PP, Onayemi A, Smyth JA, Gillan JE, Cutz E, Froese AB et al. (1983). Comparison of conventional and high-frequency ventilation: oxygenation and lung pathology. J Appl Physiol, 55, 131-138.

33 Slutsky AS, Tremblay LN (1998). Multiple system organ failure. Is mechanical ventilation a contributing factor? Am J Respir Crit Care Med, 157, 1721-1725.

34 Gattinoni L, Pesenti A, Avalli L, Rossi F, Bombino M (1987). Pressure-volume curve of total respiratory system in acute respiratory failure. Computed tomographic scan study. Am Rev Respir Dis, 136, 730-736.

35 Gattinoni L, Mascheroni D, Torresin A, Marcolin R, Fumagalli R, Vesconi S et al. (1986). Morphological response to positive end expiratory pressure in acute respiratory failure. Computerized tomography study. Intensive Care Med, 12, 137-142.

36 Hickling KG, Henderson SJ, Jackson R (1990). Low mortality associated with low volume pressure limited ventilation with permissive hypercapnia in severe adult respiratory distress syndrome. Intensive Care Med, 16, 372-377.

37 Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network (2000). N Engl J Med, 342, 1301-1308.

38 Amato MB, Barbas CS, Medeiros DM, Schettino GP, Lorenzi FG, Kairalla RA et al. (1995). Beneficial effects of the "open lung approach" with low distending pressures in acute respiratory distress syndrome. A prospective randomized study on mechanical ventilation. Am J Respir Crit Care Med, 152, 1835-1846.

39 Amato MB, Barbas CS, Medeiros DM, Magaldi RB, Schettino GP, Lorenzi-Filho G et al. (1998). Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med, 338, 347-354.

40 Brochard L, Roudot-Thoraval F, Roupie E, Delclaux C, Chastre J, Fernandez-Mondejar E et al. (1998). Tidal volume reduction for prevention of ventilator-induced lung injury in acute respiratory distress syndrome. The Multicenter Trail Group on Tidal Volume reduction in ARDS. Am J Respir Crit Care Med, 158, 1831-1838.

41 Stewart TE, Meade MO, Cook DJ, Granton JT, Hodder RV, Lapinsky SE et al. (1998). Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. Pressure- and Volume-Limited Ventilation Strategy Group. N Engl J Med, 338, 355-361.

Page 135: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

134

42 Ranieri VM, Suter PM, Tortorella C, De TR, Dayer JM, Brienza A et al. (1999). Effect of mechanical ventilation on inflammatory mediators in patients with acute respiratory distress syndrome: a randomized controlled trial. JAMA, 282, 54-61.

43 Froese AB (1997). High-frequency oscillatory ventilation for adult respiratory distress syndrome: let's get it right this time! Crit Care Med, 25, 906-908.

44 Lachmann B (1992). Open up the lung and keep the lung open. Intensive Care Med, 18, 319-321.

45 Markhorst DG, van Genderingen HR, van Vught AJ (2004). Static pressure-volume curve characteristics are moderate estimators of optimal airway pressures in a mathematical model of (primary/pulmonary) acute respiratory distress syndrome. Intensive Care Med, 30, 2086-2093.

46 Curley MA, Hibberd PL, Fineman LD, Wypij D, Shih MC, Thompson JE et al. (2005). Effect of prone positioning on clinical outcomes in children with acute lung injury: a randomized controlled trial. JAMA, 294, 229-237.

47 Sokol J, Jacobs SE, Bohn D (2003). Inhaled nitric oxide for acute hypoxic respiratory failure in children and adults: a meta-analysis. Anesth Analg, 97, 989-998.

48 Spear RM, Fackler JC (1998). Extracorporeal membrane oxygenation and pediatric acute respiratory distress syndrome: we can afford it, but we don't need it. Crit Care Med, 26, 1486-1487.

49 Willson DF, Thomas NJ, Markovitz BP, Bauman LA, DiCarlo JV, Pon S et al. (2005). Effect of exogenous surfactant (calfactant) in pediatric acute lung injury: a randomized controlled trial. JAMA, 293, 470-476.

50 Imai Y, Slutsky AS (2005). High-frequency oscillatory ventilation and ventilator-induced lung injury. Crit Care Med, 33, S129-S134.

51 Meyer J, Cox PN, McKerlie C, Bienzle D (2006). Protective strategies of high-frequency oscillatory ventilation in a rabbit model. Pediatr Res, 60, 401-406.

52 Bollen CW, van Well GT, Sherry T, Beale RJ, Shah S, Findlay G et al. (2005). High frequency oscillatory ventilation compared with conventional mechanical ventilation in adult respiratory distress syndrome: a randomized controlled trial. Crit Care, 9, R430-R439.

53 Fort P, Farmer C, Westerman J, Johannigman J, Beninati W, Dolan S et al. (1997). High-frequency oscillatory ventilation for adult respiratory distress syndrome - a pilot study. Crit Care Med, 25, 937-947.

54 Mehta S, Lapinsky SE, Hallett DC, Merker D, Groll RJ, Cooper AB et al. (2001). Prospective trial of high-frequency oscillation in adults with acute respiratory distress syndrome. Crit Care Med, 29, 1360-1369.

55 Andersen FA, Guttormsen AB, Flaatten HK (2002). High frequency oscillatory ventilation in adult patients with acute respiratory distress syndrome - a retrospective study. Acta Anaesthesiol Scand, 46, 1082-1088.

56 Cartotto R, Ellis S, Gomez M, Cooper A, Smith T (2004). High frequency oscillatory ventilation in burn patients with the acute respiratory distress syndrome. Burns, 30, 453-463.

57 Claridge JA, Hostetter RG, Lowson SM, Young JS (1999). High-frequency oscillatory ventilation can be effective as rescue therapy for refractory acute lung dysfunction. Am Surg, 65, 1092-1096.

58 David M, Weiler N, Heinrichs W, Neumann M, Joost T, Markstaller K et al. (2003). High-frequency oscillatory ventilation in adult acute respiratory distress syndrome. Intensive Care Med, 29, 1656-1665.

59 David M, Karmrodt J, Weiler N, Scholz A, Markstaller K, Eberle B (2005). High-frequency oscillatory ventilation in adults with traumatic brain injury and acute respiratory distress syndrome. Acta Anaesthesiol Scand, 49, 209-214.

60 Kao KC, Tsai YH, Wu YK, Huang CT, Shih MJ, Huang CC (2006). High frequency oscillatory ventilation for surgical patients with acute respiratory distress syndrome. J Trauma, 61, 837-843.

Page 136: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

135

61 Ferguson ND, Stewart TE (2002). New therapies for adults with acute lung injury. High-frequency oscillatory ventilation. Crit Care Clin, 18, 91-106.

62 Mehta S, Granton J, MacDonald RJ, Bowman D, Matte-Martyn A, Bachman T et al. (2004). High-frequency oscillatory ventilation in adults: the Toronto experience. Chest, 126, 518-527.

63 Ferguson ND, Chiche JD, Kacmarek RM, Hallett DC, Mehta S, Findlay GP et al. (2005). Combining high-frequency oscillatory ventilation and recruitment maneuvers in adults with early acute respiratory distress syndrome: the Treatment with Oscillation and an Open Lung Strategy (TOOLS) Trial pilot study. Crit Care Med, 33, 479-486.

64 Pillow JJ (2005). High-frequency oscillatory ventilation: mechanisms of gas exchange and lung mechanics. Crit Care Med, 33, S135-S141.

65 Venegas JG, Fredberg JJ (1994). Understanding the pressure cost of ventilation: why does high-frequency ventilation work? Crit Care Med, 22, S49-S57.

66 Neumann P, Wrigge H, Zinserling J, Hinz J, Maripuu E, Andersson LG et al. (2005). Spontaneous breathing affects the spatial ventilation and perfusion distribution during mechanical ventilatory support. Crit Care Med, 33, 1090-1095.

67 Putensen C, Mutz NJ, Putensen-Himmer G, Zinserling J (1999). Spontaneous breathing during ventilatory support improves ventilation-perfusion distributions in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med, 159, 1241-1248.

68 Wrigge H, Zinserling J, Neumann P, Defosse J, Magnusson A, Putensen C et al. (2003). Spontaneous breathing improves lung aeration in oleic acid-induced lung injury. Anesthesiology, 99, 376-384.

69 Wrigge H, Zinserling J, Neumann P, Muders T, Magnusson A, Putensen C et al. (2005). Spontaneous breathing with airway pressure release ventilation favors ventilation in dependent lung regions and counters cyclic alveolar collapse in oleic-acid-induced lung injury: a randomized controlled computed tomography trial. Crit Care, 9, R780-R789.

70 Hering R, Peters D, Zinserling J, Wrigge H, von Spiegel T, Putensen C (2002). Effects of spontaneous breathing during airway pressure release ventilation on renal perfusion and function in patients with acute lung injury. Intensive Care Med, 28, 1426-1433.

71 Varpula T, Valta P, Niemi R, Takkunen O, Hynynen M, Pettila VV (2004). Airway pressure release ventilation as a primary ventilatory mode in acute respiratory distress syndrome. Acta Anaesthesiol Scand, 48, 722-731.

72 Forel JM, Roch A, Marin V, Michelet P, Demory D, Blache JL et al. (2006). Neuromuscular blocking agents decrease inflammatory response in patients presenting with acute respiratory distress syndrome. Crit Care Med, 34, 2749-2757.

73 Banner MJ, Jaeger MJ, Kirby RR (1994). Components of the work of breathing and implications for monitoring ventilator-dependent patients. Crit Care Med, 22, 515-523.

74 Kirton OC, DeHaven CB, Morgan JP, Windsor J, Civetta JM (1995). Elevated imposed work of breathing masquerading as ventilator weaning intolerance. Chest, 108, 1021-1025.

75 Campbell EJ. (1958). The respiratory muscles and the mechanics of breathing. Chicago, Year Book Medical.

76 Sassoon CS, Light RW, Lodia R, Sieck GC, Mahutte CK (1991). Pressure-time product during continuous positive airway pressure, pressure support ventilation, and T-piece during weaning from mechanical ventilation. Am Rev Respir Dis, 143, 469-475.

77 Collett PW, Perry C, Engel LA (1985). Pressure-time product, flow, and oxygen cost of resistive breathing in humans. J Appl Physiol, 58, 1263-1272.

78 Chiumello D, Pelosi P, Croci M, Bigatello LM, Gattinoni L (2001). The effects of pressurization rate on breathing pattern, work of breathing, gas exchange and patient comfort in pressure support ventilation. Eur Respir J, 18, 107-114.

Page 137: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

136

79 Vitacca M, Bianchi L, Zanotti E, Vianello A, Barbano L, Porta R et al. (2004). Assessment of physiologic variables and subjective comfort under different levels of pressure support ventilation. Chest, 126, 851-859.

80 French CJ (1999). Work of breathing measurement in the critically ill patient. Anaesth Intensive Care, 27, 561-573.

81 Girault C, Breton L, Richard JC, Tamion F, Vandelet P, Aboab J et al. (2003). Mechanical effects of airway humidification devices in difficult to wean patients. Crit Care Med, 31, 1306-1311.

82 MacIntyre NR (2005). Respiratory mechanics in the patient who is weaning from the ventilator. Respir Care, 50, 275-286.

83 MacIntyre NR, Cook DJ, Ely EW, Jr., Epstein SK, Fink JB, Heffner JE et al. (2001). Evidence-based guidelines for weaning and discontinuing ventilatory support: a collective task force facilitated by the American College of Chest Physicians; the American Association for Respiratory Care; and the American College of Critical Care Medicine. Chest, 120, 375S-395S.

84 Putensen C, Hering R, Muders T, Wrigge H (2005). Assisted breathing is better in acute respiratory failure. Curr Opin Crit Care, 11, 63-68.

85 Zapletal A, Samanek M, Paul T (1987). Lung Function in Children and Adolescents. Basel: Karger.

86 Thibeault DW, Clutario B, Awld PA (1966). The oxygen cost of breathing in the premature infant. Pediatrics, 37, 954-959.

87 Butler WJ, Bohn DJ, Bryan AC, Froese AB (1980). Ventilation by high-frequency oscillation in humans. Anesth Analg, 59, 577-584.

88 DeWeese EL, Sullivan TY, Yu PL (1985). Ventilatory response to high-frequency airway oscillation in humans. J Appl Physiol, 58, 1099-1106.

89 England SJ, Sullivan C, Bowes G, Onayemi A, Bryan AC (1985). State-related incidence of spontaneous breathing during high frequency ventilation. Respir Physiol, 60, 357-364.

90 George RJ, Winter RJ, Johnson MA, Slee IP, Geddes DM (1985). Effect of oral high frequency ventilation by jet or oscillator on minute ventilation in normal subjects. Thorax, 40, 749-755.

91 Thompson-Gorman SL, Fitzgerald RS, Mitzner W (1988). Chemical and mechanical determinants of apnea during high-frequency ventilation. J Appl Physiol, 65, 179-186.

92 van Vught AJ, Versprille A, Jansen JR (1986). Suppression of spontaneous breathing during high-frequency jet ventilation. Influence of dynamic changes and static levels of lung stretch. Intensive Care Med, 12, 26-32.

93 van Vught AJ, Versprille A, Jansen JR (1987). Suppression of spontaneous breathing during high-frequency jet ventilation. Separate effects of lung volume and jet frequency. Intensive Care Med, 13, 315-322.

94 Sessler CN (2005). Sedation, analgesia, and neuromuscular blockade for high-frequency oscillatory ventilation. Crit Care Med, 33, S209-S216.

95 Cereda M, Foti G, Marcora B, Gili M, Giacomini M, Sparacino ME et al. (2000). Pressure support ventilation in patients with acute lung injury. Crit Care Med, 28, 1269-1275.

96 Putensen C, Zech S, Wrigge H, Zinserling J, Stuber F, von Spiegel T et al. (2001). Long-term effects of spontaneous breathing during ventilatory support in patients with acute lung injury. Am J Respir Crit Care Med, 164, 43-49.

97 Sydow M, Burchardi H, Ephraim E, Zielmann S, Crozier TA (1994). Long-term effects of two different ventilatory modes on oxygenation in acute lung injury. Comparison of airway pressure release ventilation and volume-controlled inverse ratio ventilation. Am J Respir Crit Care Med, 149, 1550-1556.

98 Bollen CW, Uiterwaal CS, van Vught AJ (2003). Cumulative metaanalysis of high-frequency versus conventional ventilation in premature neonates. Am J Respir Crit Care Med, 168, 1150-1155.

Page 138: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

137

99 Froese AB, Kinsella JP (2005). High-frequency oscillatory ventilation: lessons from the neonatal/pediatric experience. Crit Care Med, 33, S115-S121.

100 Bollen CW, van Well GT, Sherry T, Beale RJ, Shah S, Findlay G et al. (2005). High frequency oscillatory ventilation compared with conventional mechanical ventilation in adult respiratory distress syndrome: a randomized controlled trial. Crit Care, 9, R430-R439.

101 Derdak S, Mehta S, Stewart TE, Smith T, Rogers M, Buchman TG et al. (2002). High-frequency oscillatory ventilation for acute respiratory distress syndrome in adults: a randomized, controlled trial. Am J Respir Crit Care Med, 166, 801-808.

102 Singh JM, Ferguson ND (2005). Is it time to increase the frequency of use of high-frequency oscillatory ventilation? Crit Care, 9, 339-340.

103 Slee-Wijffels FY, van der Vaart KR, Twisk JW, Markhorst DG, Plötz FB (2005). High-frequency oscillatory ventilation in children: a single-center experience of 53 cases. Crit Care, 9, R274-R279.

104 Kacmarek RM, Hess D (1994). Basic principles of ventilator machinery. In: JW Twisk (Ed), Principles and practice of mechanical ventilation, pp. 65-110. McGraw-Hill, Inc.

105 SensorMedics:3100 A High Frequency Oscillatory Ventilator Operator's Manual (1993). Yorba Linda, CA: SensorMedics.

106 SensorMedics:3100 B High Frequency Oscillatory Ventilator Operator's Manual (1999). Yorba Linda, CA: SensorMedics.

107 Agostini E, Campbell EJ, Freedman S (1970). Energetics. In: Campbell EJ, Agostini E, Newsom Davis J (Eds), The respiratory muscles, pp. 115-137. Philadelphia, Saunders.

108 Putensen C, Muders T, Varelmann D, Wrigge H (2006). The impact of spontaneous breathing during mechanical ventilation. Curr Opin Crit Care, 12, 13-18.

109 Chatburn RL (2004). Computer control of mechanical ventilation. Respir Care, 49, 507-517.

110 Casaseca-de-la-Higuera P, Simmross-Wattenberg F, Martin-Fernandez M, berola-Lopez C (2009). A multichannel model-based methodology for extubation readiness decision of patients on weaning trials. IEEE Trans Biomed Eng, 56, 1849-1863.

111 Jandre FC, Pino AV, Lacorte I, Neves JH, Giannella-Neto A (2004). A closed-loop mechanical ventilation controller with explicit objective functions. IEEE Trans Biomed Eng, 51, 823-831.

112 van Heerde M, van Genderingen HR, Leenhoven T, Roubik K, Plötz FB, Markhorst DG (2006). Imposed work of breathing during high-frequency oscillatory ventilation: a bench study. Crit Care, 10, R23.

113 Fessler HE, Derdak S, Ferguson ND, Hager DN, Kacmarek RM, Thompson BT et al. (2007). A protocol for high-frequency oscillatory ventilation in adults: results from a roundtable discussion. Crit Care Med, 35, 1649-1654.

114 Kopelent, V. (2007). Artificial lung ventilation and its optimization. Thesis. Czech Technical University in Prague, Kladno, Czech Republic.

115 Åström KJ, Wittenmark B (2007). Computer controlled systems theory and design. (3rd ed) Englewood Cliffs, N.J: Prentice-Hall.

116 Anderson BD, Moore JB (2007). Optimal Control: Linear Quadratic Methods . Mineola, NY: Dover Publications.

117 Simon D (2006). Optimal state estimation Kalman, H[infinity], and nonlinear approaches. Hoboken, N.J: Wiley-Interscience.

118 Jiao GY, Newhart JW (2008). Bench study on active exhalation valve performance. Respir Care, 53, 1697-1702.

119 Mireles-Cabodevila E, Chatburn RL (2009). Work of breathing in adaptive pressure control continuous mandatory ventilation. Respir Care, 54, 1467-1472.

120 Terado M, Ichiba S, Nagano O, Ujike Y (2008). Evaluation of pressure support ventilation with seven different ventilators using Active Servo Lung 5000. Acta Med Okayama, 62, 127-133.

Page 139: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

138

121 Yi, W., Zhang, Q., Wang, Y., & Qin, H. (2008). On-line non-invasive determination of patients and ventilator respiratory work during Proportional Assist Ventilation. Zhanggjiajie, China (pp. 1163-1167).

122 Thille AW, Lyazidi A, Richard JC, Galia F, Brochard L (2009). A bench study of intensive-care-unit ventilators: new versus old and turbine-based versus compressed gas-based ventilators. Intensive Care Med, 35, 1368-1376.

123 Gerstmann DR, Fouke JM, Winter DC, Taylor AF, deLemos RA (1990). Proximal, tracheal, and alveolar pressures during high-frequency oscillatory ventilation in a normal rabbit model. Pediatr Res, 28, 367-373.

124 Derdak S (2005). High-frequency oscillatory ventilation for adult acute respiratory distress syndrome: a decade of progress. Crit Care Med, 33, S113-S114.

125 Banner MJ, Downs JB, Kirby RR, Smith RA, Boysen PG, Lampotang S (1988). Effects of expiratory flow resistance on inspiratory work of breathing. Chest, 93, 795-799.

126 van Heerde M, Roubik K, Kopelent V, Plötz FB, Markhorst DG (2006). Unloading work of breathing during high-frequency oscillatory ventilation: a bench study. Crit Care, 10, R103.

127 Grotjohan, H. & van der Heijde, R. (2006). Experimental models of the respiratory distress syndrome: lavage and oleic acid. Thesis. Erasmus University Rotterdam, Rotterdam, the Netherlands

128 Lachmann B, Robertson B, Vogel J (1980). In vivo lung lavage as an experimental model of the respiratory distress syndrome. Acta Anaesthesiol Scand, 24, 231-236.

129 Katz JA, Kraemer RW, Gjerde GE (1985). Inspiratory work and airway pressure with continuous positive airway pressure delivery systems. Chest, 88, 519-526.

130 Baydur A, Behrakis PK, Zin WA, Jaeger M, Milic-Emili J (1982). A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis, 126, 788-791.

131 Cabello B, Mancebo J (2006). Work of breathing. Intensive Care Med, 32, 1311-1314.

132 Jubran A, Van de Graaff WB, Tobin MJ (1995). Variability of patient-ventilator interaction with pressure support ventilation in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med, 152, 129-136.

133 Guyton AC, Hall JE (2006). Textbook of Medical Physiology. (11 ed) Philadelphia: Elsevier Saunders.

134. Tingay DG, Mills JF, Morley CJ, Pellicano A, Dargaville PA (2006). The deflation limb of the pressure-volume relationship in infants during high-frequency ventilation. Am J Respir Crit Care Med, 173, 414-420.

135 Otis AB, Fenn WO, Rahn H (1950). Mechanics of breathing in man. J Appl Physiol, 2, 592-607.

136 Schultz MJ, Haitsma JJ, Slutsky AS, Gajic O (2007). What tidal volumes should be used in patients without acute lung injury? Anesthesiology, 106, 1226-1231.

137 Thille AW, Rodriguez P, Cabello B, Lellouche F, Brochard L (2006). Patient-ventilator asynchrony during assisted mechanical ventilation. Intensive Care Med, 32, 1515-1522.

138 Muellenbach RM, Kredel M, Said HM, Klosterhalfen B, Zollhoefer B, Wunder C et al. (2007). High-frequency oscillatory ventilation reduces lung inflammation: a large-animal 24-h model of respiratory distress. Intensive Care Med, 33, 1423-1433.

139 Neumann PF, Schubert AF, Heuer JF, Hinz JF, Quintel MF, Klockgether-Radke A Hemodynamic effects of spontaneous breathing in the post-operative period.

140 Varelmann D, Muders T, Zinserling J, Guenther U, Magnusson A, Hedenstierna G et al. (2008). Cardiorespiratory effects of spontaneous breathing in two different models of experimental lung injury: a randomized controlled trial. Crit Care, 12, R135.

141 Derdak S (2003). High-frequency oscillatory ventilation for acute respiratory distress syndrome in adult patients. Crit Care Med, 31, S317-S323.

142 Higgins J, Estetter B, Holland D, Smith B, Derdak S (2005). High-frequency oscillatory ventilation in adults: respiratory therapy issues. Crit Care Med, 33, S196-S203.

Page 140: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

139

143 van Heerde M, Roubik K, Kopelent V, Plötz FB, Markhorst DG (2009). Demand flow facilitates spontaneous breathing during high-frequency oscillatory ventilation in a pig model. Crit Care Med, 37, 1068-1073.

144 Richard JC, Pouzot C, Gros A, Tourevieille C, Lebars D, Lavenne F et al. (2009). Electrical impedance tomography compared to positron emission tomography for the measurement of regional lung ventilation: an experimental study. Crit Care, 13, R82.

145 Hinz J, Moerer O, Neumann P, Dudykevych T, Frerichs I, Hellige G et al. (2006). Regional pulmonary pressure volume curves in mechanically ventilated patients with acute respiratory failure measured by electrical impedance tomography. Acta Anaesthesiol Scand, 50, 331-339.

146 Schibler A, Calzia E (2008). Electrical impedance tomography: a future item on the "Christmas Wish List" of the intensivist? Intensive Care Med, 34, 400-401.

147 Brown BH (2003). Electrical impedance tomography (EIT): a review. J Med Eng Technol, 27, 97-108.

148 Wolf GK, Arnold JH (2005). Noninvasive assessment of lung volume: respiratory inductance plethysmography and electrical impedance tomography. Crit Care Med, 33, S163-S169.

149 Hahn G, Frerichs I, Kleyer M, Hellige G (1996). Local mechanics of the lung tissue determined by functional EIT. Physiol Meas, 17 Suppl 4A, A159-A166.

150 van Genderingen HR, van Vught AJ, Jansen JR (2003). Estimation of regional lung volume changes by electrical impedance pressures tomography during a pressure-volume maneuver. Intensive Care Med, 29, 233-240.

151 Frerichs I, Dargaville PA, van GH, Morel DR, Rimensberger PC (2006). Lung volume recruitment after surfactant administration modifies spatial distribution of ventilation. Am J Respir Crit Care Med, 174, 772-779.

152 Frerichs I, Dudykevych T, Hinz J, Bodenstein M, Hahn G, Hellige G (2001). Gravity effects on regional lung ventilation determined by functional EIT during parabolic flights. J Appl Physiol, 91, 39-50.

153 Hinz J, Gehoff A, Moerer O, Frerichs I, Hahn G, Hellige G et al. (2007). Regional filling characteristics of the lungs in mechanically ventilated patients with acute lung injury. Eur J Anaesthesiol, 24, 414-424.

154 Seymour CW, Frazer M, Reilly PM, Fuchs BD (2007). Airway pressure release and biphasic intermittent positive airway pressure ventilation: are they ready for prime time? J Trauma, 62, 1298-1308.

155 Gonzalez M, Arroliga AC, Frutos-Vivar F, Raymondos K, Esteban A, Putensen C et al. (2010). Airway pressure release ventilation versus assist-control ventilation: a comparative propensity score and international cohort study. Intensive Care Med.

156 Forel JM, Roch A, Papazian L (2009). Paralytics in critical care: not always the bad guy. Curr Opin Crit Care, 15, 59-66.

157 Papazian L, Forel JM, Gacouin A, Penot-Ragon C, Perrin G, Loundou A et al. (2010). Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med, 363, 1107-1116.

158 Esan A, Hess DR, Raoof S, George L, Sessler CN (2010). Severe hypoxemic respiratory failure: part 1 - ventilatory strategies. Chest, 137, 1203-1216.

159 Aspros AJ, Coto CG, Lewis JF, Veldhuizen RA (2010). High-frequency oscillation and surfactant treatment in an acid aspiration model. Can J Physiol Pharmacol, 88, 14-20.

160 Bohn DJ, Miyasaka K, Marchak BE, Thompson WK, Froese AB, Bryan AC (1980). Ventilation by high-frequency oscillation. J Appl Physiol, 48, 710-716.

Page 141: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

140

161 der Hardt K, Kandler MA, Fink L, Schoof E, Dotsch J, Brandenstein O et al. (2004). High frequency oscillatory ventilation suppresses inflammatory response in lung tissue and microdissected alveolar macrophages in surfactant depleted piglets. Pediatr Res, 55, 339-346.

162 Jian MY, Koizumi T, Yokoyama T, Tsushima K, Kubo K (2010). Comparison of acid-induced inflammatory responses in the rat lung during high frequency oscillatory and conventional mechanical ventilation. Inflamm Res.

163 McCulloch PR, Forkert PG, Froese AB (1988). Lung volume maintenance prevents lung injury during high frequency oscillatory ventilation in surfactant-deficient rabbits. Am Rev Respir Dis, 137, 1185-1192.

164 Nakagawa R, Koizumi T, Ono K, Yoshikawa S, Tsushima K, Otagiri T (2008). Effects of high-frequency oscillatory ventilation on oleic acid-induced lung injury in sheep. Lung, 186, 225-232.

165 Rotta AT, Gunnarsson B, Fuhrman BP, Hernan LJ, Steinhorn DM (2001). Comparison of lung protective ventilation strategies in a rabbit model of acute lung injury. Crit Care Med, 29, 2176-2184.

166 Shimaoka M, Fujino Y, Taenaka N, Hiroi T, Kiyono H, Yoshiya I, I (1998). High frequency oscillatory ventilation attenuates the activation of alveolar macrophages and neutrophils in lung injury. Crit Care, 2, 35-39.

167 Slutsky AS, Drazen FM, Ingram RH, Jr., Kamm RD, Shapiro AH, Fredberg JJ et al. (1980). Effective pulmonary ventilation with small-volume oscillations at high frequency. Science, 209, 609-671.

168 Sugiura M, McCulloch PR, Wren S, Dawson RH, Froese AB (1994). Ventilator pattern influences neutrophil influx and activation in atelectasis-prone rabbit lung. J Appl Physiol, 77, 1355-1365.

169 Takata M, Abe J, Tanaka H, Kitano Y, Doi S, Kohsaka T et al. (1997). Intraalveolar expression of tumor necrosis factor-alpha gene during conventional and high-frequency ventilation. Am J Respir Crit Care Med, 156, 272-279.

170 Imai Y, Nakagawa S, Ito Y, Kawano T, Slutsky AS, Miyasaka K (2001). Comparison of lung protection strategies using conventional and high-frequency oscillatory ventilation. J Appl Physiol, 91, 1836-1844.

171 Carney D, DiRocco J, Nieman G (2005). Dynamic alveolar mechanics and ventilator-induced lung injury. Crit Care Med, 33, S122-S128.

172 Boynton BR, Hammond MD, Fredberg JJ, Buckley BG, Villanueva D, Frantz ID, III (1989). Gas exchange in healthy rabbits during high-frequency oscillatory ventilation. J Appl Physiol, 66, 1343-1351.

173 Kamitsuka MD, Boynton BR, Villanueva D, Vreeland PN, Frantz ID, III (1990). Frequency, tidal volume, and mean airway pressure combinations that provide adequate gas exchange and low alveolar pressure during high frequency oscillatory ventilation in rabbits. Pediatr Res, 27, 64-69.

174 Schindler M, Seear M (1991). The effect of lung mechanics on gas transport during high-frequency oscillation. Pediatr Pulmonol, 11, 335-339.

175 Slutsky AS, Kamm RD, Rossing TH, Loring SH, Lehr J, Shapiro AH et al. (1981). Effects of frequency, tidal volume, and lung volume on CO2 elimination in dogs by high frequency (2-30 Hz), low tidal volume ventilation. J Clin Invest, 68, 1475-1484.

176 Erickson S, Schibler A, Numa A, Nuthall G, Yung M, Pascoe E et al. (2007). Acute lung injury in pediatric intensive care in Australia and New Zealand: a prospective, multicenter, observational study. Pediatr Crit Care Med, 8, 317-323.

177 Randolph AG, Meert KL, O'Neil ME, Hanson JH, Luckett PM, Arnold JH et al. (2003). The feasibility of conducting clinical trials in infants and children with acute respiratory failure. Am J Respir Crit Care Med, 167, 1334-1340.

178 Santschi M, Jouvet P, Leclerc F, Gauvin F, Newth CJ, Carroll CL et al. (2010). Acute lung injury in children: Therapeutic practice and feasibility of international clinical trials. Pediatr Crit Care Med.

Page 142: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

141

179 Zimmerman JJ, Akhtar SR, Caldwell E, Rubenfeld GD (2009). Incidence and outcomes of pediatric acute lung injury. Pediatrics, 124, 87-95.

180 Arnold JH, Hanson JH, Toro-Figuero LO, Gutierrez J, Berens RJ, Anglin DL (1994). Prospective, randomized comparison of high-frequency oscillatory ventilation and conventional mechanical ventilation in pediatric respiratory failure. Crit Care Med, 22, 1530-1539.

181 Samransamruajkit R, Prapphal N, Deelodegenavong J, Poovorawan Y (2005). Plasma soluble intercellular adhesion molecule-1 (sICAM-1) in pediatric ARDS during high frequency oscillatory ventilation: a predictor of mortality. Asian Pac J Allergy Immunol, 23, 181-188.

182 Lochindarat S, Srisan P, Jatanachai P (2003). Factors effecting the outcome of acute respiratory distress syndrome in pediatric patients treated with high frequency oscillatory ventilation. J Med Assoc Thai, 86 Suppl 3, S618-S627.

183 Watkins SJ, Peters MJ, Tasker RC (2000). One hundred courses of high frequency oscillatory ventilation: what have we learned? Eur J Pediatr, 159, 134.

184 Sarnaik AP, Meert KL, Pappas MD, Simpson PM, Lieh-Lai MW, Heidemann SM (1996). Predicting outcome in children with severe acute respiratory failure treated with high-frequency ventilation. Crit Care Med, 24, 1396-1402.

185 Berner ME, Hanquinet S, Rimensberger PC (2008). High frequency oscillatory ventilation for respiratory failure due to RSV bronchiolitis. Intensive Care Med, 34, 1698-1702.

186 Arnold JH, Anas NG, Luckett P, Cheifetz IM, Reyes G, Newth CJ et al. (2000). High-frequency oscillatory ventilation in pediatric respiratory failure: a multicenter experience. Crit Care Med, 28, 3913-3919.

187 Brogan TV, Bratton SL, Meyer RJ, O'Rourke PP, Jardine DS (2000). Nonpulmonary organ failure and outcome in children treated with high-frequency oscillatory ventilation. J Crit Care, 15, 5-11.

188 Martinon Torres F, Rodriguez Nunez A, Jaimovich DG, Martinon Sanchez JM (2000). High-frequency oscillatory ventilation in pediatric patients. protocol and preliminary results. An Esp Pediatr, 53, 305-313.

189 Ben Jaballah N, Khaldi A, Mnif K, Bouziri A, Belhadj S, Hamdi A et al. (2006). High-frequency oscillatory ventilation in pediatric patients with acute respiratory failure. Pediatr Crit Care Med, 7, 362-367.

190 Duval EL, Markhorst DG, Gemke RJ, van Vught AJ (2000). High-frequency oscillatory ventilation in pediatric patients. Neth J Med, 56, 177-185.

191 Anton N, Joffe KM, Joffe AR (2003). Inability to predict outcome of acute respiratory distress syndrome in children when using high frequency oscillation. Intensive Care Med, 29, 1763-1769.

192 Rosenberg RB, Broner CW, Peters KJ, Anglin DL (1993). High-frequency ventilation for acute pediatric respiratory failure. Chest, 104, 1216-1221.

193 Fedora M, Klimovic M, Seda M, Dominik P, Nekvasil R (2000). Effect of early intervention of high-frequency oscillatory ventilation on the outcome in pediatric acute respiratory distress syndrome. Bratisl Lek Listy, 101, 8-13.

194 Dobyns EL, Anas NG, Fortenberry JD, Deshpande J, Cornfield DN, Tasker RC et al. (2002). Interactive effects of high-frequency oscillatory ventilation and inhaled nitric oxide in acute hypoxemic respiratory failure in pediatrics. Crit Care Med, 30, 2425-2429.

195 Leipala JA, Sharma A, Lee S, Milner AD, Greenough A (2005). An in vitro assessment of gas trapping during high frequency oscillation. Physiol Meas, 26, 329-336.

196 David M, Markstaller K, Depta AL, Karmrodt J, Herweling A, Kempski O et al. (2006). Initiation of high-frequency oscillatory ventilation and its effects upon cerebral circulation in pigs: an experimental study. Br J Anaesth, 97, 525-532.

Page 143: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

142

197 O'Rourke J, Sheeran P, Heaney M, Talbot R, Geraghty M, Costello J et al. (2007). Effects of sequential changes from conventional ventilation to high-frequency oscillatory ventilation at increasing mean airway pressures in an ovine model of combined lung and head injury. Eur J Anaesthesiol, 24, 454-463.

198 Suzuki H, Papazoglou K, Bryan AC (1992). Relationship between PaO2 and lung volume during high frequency oscillatory ventilation. Acta Paediatr Jpn, 34, 494-500.

199 Bond DM, Froese AB (1993). Volume recruitment maneuvers are less deleterious than persistent low lung volumes in the atelectasis-prone rabbit lung during high-frequency oscillation. Crit Care Med, 21, 402-412.

200 Bond DM, McAloon J, Froese AB (1994). Sustained inflations improve respiratory compliance during high-frequency oscillatory ventilation but not during large tidal volume positive-pressure ventilation in rabbits. Crit Care Med, 22, 1269-1277.

201 Sakai T, Kakizawa H, Aiba S, Takahashi R, Yoshioka T, Iinuma K (1999). Effects of mean and swing pressures on piston-type high-frequency oscillatory ventilation in rabbits with and without acute lung injury. Pediatr Pulmonol, 27, 328-335.

202 Gattinoni L, Caironi P, Cressoni M, Chiumello D, Ranieri VM, Quintel M et al. (2006). Lung recruitment in patients with the acute respiratory distress syndrome. N Engl J Med, 354, 1775-1786.

203 Pellicano A, Tingay DG, Mills JF, Fasulakis S, Morley CJ, Dargaville PA (2009). Comparison of four methods of lung volume recruitment during high frequency oscillatory ventilation. Intensive Care Med, 35, 1990-1998.

204 Boynton BR, Villanueva D, Hammond MD, Vreeland PN, Buckley B, Frantz ID, III (1991). Effect of mean airway pressure on gas exchange during high-frequency oscillatory ventilation. J Appl Physiol, 70, 701-707.

205 Goddon S, Fujino Y, Hromi JM, Kacmarek RM (2001). Optimal mean airway pressure during high-frequency oscillation: predicted by the pressure- volume curve. Anesthesiology, 94, 862-869.

206 Kacmarek RM, Malhotra A (2005). High-frequency oscillatory ventilation: what large-animal studies have taught us! Crit Care Med, 33, S148-S154.

207 Luecke T, Meinhardt JP, Herrmann P, Weisser G, Pelosi P, Quintel M (2003). Setting mean airway pressure during high-frequency oscillatory ventilation according to the static pressure--volume curve in surfactant-deficient lung injury: a computed tomography study. Anesthesiology, 99, 1313-1322.

208 Chan KP, Stewart TE (2005). Clinical use of high-frequency oscillatory ventilation in adult patients with acute respiratory distress syndrome. Crit Care Med, 33, S170-S174.

209 Pillow JJ, Neil H, Wilkinson MH, Ramsden CA (1999). Effect of I/E ratio on mean alveolar pressure during high-frequency oscillatory ventilation. J Appl Physiol, 87, 407-414.

210 Kneyber MC, Plötz FB, Sibarani-Ponsen RD, Markhorst DG (2005). High-frequency oscillatory ventilation (HFOV) facilitates CO2 elimination in small airway disease: the open airway concept. Respir Med, 99, 1459-1461.

211 Pillow JJ, Sly PD, Hantos Z, Bates JH (2002). Dependence of intrapulmonary pressure amplitudes on respiratory mechanics during high-frequency oscillatory ventilation in preterm lambs. Pediatr Res, 52, 538-544.

212 van Genderingen HR, Versprille A, Leenhoven T, Markhorst DG, van Vught AJ, Heethaar RM (2001). Reduction of oscillatory pressure along the endotracheal tube is indicative for maximal respiratory compliance during high-frequency oscillatory ventilation: a mathematical model study. Pediatr Pulmonol, 31, 458-463.

213 Hamel DS, Katz AL, Craig DM, Davies JD, Cheifetz IM (2005). Carbon dioxide elimination and gas displacement vary with piston position during high-frequency oscillatory ventilation. Respir Care, 50, 361-366.

214 Scalfaro P, Pillow JJ, Sly PD, Cotting J (2001). Reliable tidal volume estimates at the airway opening with an infant monitor during high-frequency oscillatory ventilation. Crit Care Med, 29, 1925-1930.

Page 144: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

143

215 Gavriely N, Solway J, Loring SH, Butler JP, Slutsky AS, Drazen JM (1985). Pressure-flow relationships of endotracheal tubes during high-frequency ventilation. J Appl Physiol, 59, 3-11.

216 Hager DN, Fessler HE, Kaczka DW, Shanholtz CB, Fuld MK, Simon BA et al. (2007). Tidal volume delivery during high-frequency oscillatory ventilation in adults with acute respiratory distress syndrome. Crit Care Med, 35, 1522-1529.

217 Niederer PF, Leuthold R, Bush EH, Spahn DR, Schmid ER (1994). High-frequency ventilation: oscillatory dynamics. Crit Care Med, 22, S58-S65.

218 Hirao O, Iguchi N, Uchiyama A, Mashimo T, Nishimura M, Fujino Y (2009). Influence of endotracheal tube bore on tidal volume during high frequency oscillatory ventilation: a model lung study. Med Sci Monit, 15, MT1-MT4.

219 Bauer K, Brucker C (2009). The role of ventilation frequency in airway reopening. J Biomech, 42, 1108-1113.

220 Tsuzaki K, Hales CA, Strieder DJ, Venegas JG (1993). Regional lung mechanics and gas transport in lungs with inhomogeneous compliance. J Appl Physiol, 75, 206-216.

221 Custer JW, Ahmed A, Kaczka DW, Mulraeany DG, Simon BA, Easley RB (2010). In vitro performance comparison of the Sensormedics 3100A and B high-frequency oscillatory ventilators. Pediatr Crit Care Med.

222 Van de Kieft M, Dorsey D, Morison D, Bravo L, Venticinque S, Derdak S (2005). High-frequency oscillatory ventilation: lessons learned from mechanical test lung models. Crit Care Med, 33, S142-S147.

223 Fessler HE, Hager DN, Brower RG (2008). Feasibility of very high-frequency ventilation in adults with acute respiratory distress syndrome. Crit Care Med, 36, 1043-1048.

224 Bryan AC, Slutsky AS (1986). Long volume during high frequency oscillation. Am Rev Respir Dis, 133, 928-930.

225 Habib RH, Pyon KH, Courtney SE (2002). Optimal high-frequency oscillatory ventilation settings by nonlinear lung mechanics analysis. Am J Respir Crit Care Med, 166, 950-953.

226 van Genderingen HR, van Vught AJ, Jansen JR (2004). Regional lung volume during high-frequency oscillatory ventilation by electrical impedance tomography. Crit Care Med, 32, 787-794.

227 Markhorst DG, Jansen JR, van Vught AJ, van Genderingen HR (2005). Breath-to-breath analysis of abdominal and rib cage motion in surfactant-depleted piglets during high-frequency oscillatory ventilation. Intensive Care Med, 31, 424-430.

228 van Genderingen HR, van Vught JA, Jansen JR, Duval EL, Markhorst DG, Versprille A (2002). Oxygenation index, an indicator of optimal distending pressure during high-frequency oscillatory ventilation? Intensive Care Med, 28, 1151-1156.

229 van Genderingen HR, van Vught AJ, Duval EL, Markhorst DG, Jansen JR (2002). Attenuation of pressure swings along the endotracheal tube is indicative of optimal distending pressure during high-frequency oscillatory ventilation in a model of acute lung injury. Pediatr Pulmonol, 33, 429-436.

230 van Heerde M, Roubik K, Kopelent V, Kneyber MC, Markhorst DG (2010). Spontaneous breathing during high-frequency oscillatory ventilation improves regional lung characteristics in experimental lung injury. Acta Anaesthesiol Scand, 54, 1248-1256.

231 Froese AB (2006). High-frequency Ventilation. In: MJ Tobin (Ed), Principles and practice of mechanical ventilation, 2nd ed., pp. 473-492. New York, McGraw-Hill.

232 Pelosi P, Chiumello D, Calvi E, Taccone P, Bottino N, Panigada M et al. (2001). Effects of different continuous positive airway pressure devices and periodic hyperinflations on respiratory function. Crit Care Med, 29, 1683-1689.

233 Sassoon CS, Giron AE, Ely EA, Light RW (1989). Inspiratory work of breathing on flow-by and demand-flow continuous positive airway pressure. Crit Care Med, 17, 1108-1114.

234 Takeuchi M, Williams P, Hess D, Kacmarek RM (2002). Continuous positive airway pressure in new-generation mechanical ventilators: a lung model study. Anesthesiology, 96, 162-172

Page 145: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

References

144

235 Athans M (1971). Role and Use of Stochastic Linear-Quadratic-Gaussian Problem in Control System Design. Ieee Transactions on Automatic Control, 6, 529.

236 Kalman RE (1960). A new approach to linear filtering and prediction problems . Journal of Basic Engineering, 82, 35-45.

237 Maybeck PS (1979). Stochastic models, estimation and control. New York: Academic Press.

238 Pierson DJ (2011). Patient-ventilator interaction. Respir Care, 56, 214-228.

239 Marini JJ (2011). Spontaneously regulated vs. controlled ventilation of acute lung injury/acute respiratory distress syndrome. Curr Opin Crit Care, 17, 24-29.

240 Chandra A, Coggeshall JW, Ravenscraft SA, Marini JJ (1994). Hyperpnea limits the volume recruited by positive end-expiratory pressure. Am J Respir Crit Care Med, 150, 911-917.

241 Coggeshall JW, Marini JJ, Newman JH (1985). Improved oxygenation after muscle relaxation in adult respiratory distress syndrome. Arch Intern Med, 145, 1718-1720.

242 Leach RM, Treacher DF (2002). The pulmonary physician in critical care 2: oxygen delivery and consumption in the critically ill. Thorax, 57, 170-177.

243 Marik PE, Kaufman D (1996). The effects of neuromuscular paralysis on systemic and splanchnic oxygen utilization in mechanically ventilated patients. Chest, 109, 1038-1042.

244 Levine S, Nguyen T, Taylor N, Friscia ME, Budak MT, Rothenberg P et al. (2008). Rapid disuse atrophy of diaphragm fibers in mechanically ventilated humans. N Engl J Med, 358, 1327-1335.

245 Gunther ML, Morandi A, Ely EW (2008). Pathophysiology of delirium in the intensive care unit. Crit Care Clin, 24, 45-65, viii.

246 Morandi A, Jackson JC, Ely EW (2009). Delirium in the intensive care unit. Int Rev Psychiatry, 21, 43-58.

247 Raghavendran K, Pryhuber GS, Chess PR, Davidson BA, Knight PR, Notter RH (2008). Pharmacotherapy of acute lung injury and acute respiratory distress syndrome. Curr Med Chem, 15, 1911-1924.

248 Marini JJ, Gattinoni L (2008). Propagation prevention: a complementary mechanism for "lung protective" ventilation in acute respiratory distress syndrome. Crit Care Med, 36, 3252-3258.

249 Strøm T, Martinussen T, Toft P (2010). A protocol of no sedation for critically ill patients receiving mechanical ventilation: a randomised trial. Lancet, 375, 475-480.

250 Kress JP, Pohlman AS, O'Connor MF, Hall JB (2000). Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med, 342, 1471-1477.

251 Ortiz G, Frutos-Vivar F, Ferguson ND, Esteban A, Raymondos K, Apezteguia C et al. (2010). Outcomes of patients ventilated with synchronized intermittent mandatory ventilation with pressure support: a comparative propensity score study. Chest, 137, 1265-1277.

252 Terzi N, Pelieu I, Guittet L, Ramakers M, Seguin A, Daubin C et al. (2010). Neurally adjusted ventilatory assist in patients recovering spontaneous breathing after acute respiratory distress syndrome: physiological evaluation. Crit Care Med, 38, 1830-1837.

Page 146: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Author’s affiliations

Dankwoord

Curriculum vitae

Page 147: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Authors’ affiliations

146

Authors’ affiliations

Huib R. van Genderingen Department of Physics and Medical Technology, VU University Medical Center, Amsterdam, the Netherlands. Marc van Heerde Department of Pediatric Intensive Care, VU University Medical Center, Amsterdam, the Netherlands. Martin C.J. Kneyber Department of Pediatric Intensive Care, Beatrix Children's Hospital, University Medical Center Groningen, The Netherlands and Department of Pediatric Intensive Care, VU University Medical Center, Amsterdam, the Netherlands. Vit Kopelent Faculty of Biomedical Engineering, Czech Technical University in Prague, Kladno, Czech Republic. Tom Leenhoven CareFusion, Houten, the Netherlands. Dick G. Markhorst Department of Pediatric Intensive Care, VU University Medical Center, Amsterdam, the Netherlands. Frans B. Plötz Department of Pediatrics, Tergooiziekenhuizen, Blaricum, the Netherlands. Jakub Ráfl Faculty of Biomedical Engineering, Czech Technical University in Prague, Kladno, Czech Republic. Karel Roubik Faculty of Biomedical Engineering, Czech Technical University in Prague, Kladno, Czech Republic.

Page 148: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Dankwoord

147

Dankwoord

Het is af! Zonder de hulp, inzet en stimulering van velen was dit proefschrift er nooit gekomen. Een aantal van hen wil ik hier met name noemen.

Dick Markhorst stond aan de basis van dit project. Jij hebt ooit bedacht dat ik ging promoveren. Door jouw vertrouwen en steun ben ik daar zelf geleidelijk ook in gaan geloven. Je positieve, maar ook kritische begeleiding heb ik enorm gewaardeerd evenals je onbeperkte vermogen om werkelijk alles te filteren. We hebben elkaar de afgelopen jaren bij diverse gelegenheden goed leren kennen, dat is veel waard gebleken.

John Roord, eerst opleider, nu promotor. Al die formules in dit boek vond je prachtig, maar lastig te vatten. Toch ben je ook voor de inhoud van dit proefschrift belangrijk geweest. Je oprechte persoonlijke belangstelling, begeleiding en steun in de afgelopen jaren worden hier zeer gewaardeerd.

The work presented in this thesis originates from a fruitful cooperation between the Faculty of Biomedical Engineering in Kadno of the Czech Technical University in Prague and the PICU of the VU University Medical Center in Amsterdam. Thanks to all Czechs participating in this project, for the effort they took to introduce me in Kalman filtering and modern control theory.

Karel Roubik, first of all thanks for your hospitality during my stays in Prague and Kladno. Without all efforts made by you and your colleagues from the FBMI there would be no thesis. I am honored to be the first international member of the wine testing committee of your faculty. I will always remember ‘the Russian evening’ in Prague and the testing of Moldavian white wine that night. I still can’t remember which of the two of us has won most Emerson Awards. Best place to meet again is Snowbird I guess.

Vit Kopelent, Vitek, you have spent many (nocturnal) hours on this project. It was a pleasure to work with you during the animal experiments and data evaluation. Thanks for your hospitality as well.

Jakub Ráfl, we actually never met in person. Your input in finishing chapter 3 is greatly acknowledged. Thank you for your contribution.

Tom Leenhoven ondersteunde dit project op alle mogelijke manieren. Jij initieerde samen met Dick dit project. Volgens de overlevering werd de basis ooit gelegd op de achterzijde van een viltje in een Brusselse bar. Jouw steun, inbreng, vermogen

Page 149: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Dankwoord

148

om mensen bij elkaar te brengen en vooral je enthousiasme waardeer ik enorm. Je bood mij de mogelijkheid om op diverse locaties in de wereld iedereen op gebied van HFOV te ontmoeten. Dat was iedere keer enorm stimulerend.

Vanuit Viasys (nu CareFusion) hebben naast Tom vele anderen hun bijdrage geleverd aan dit proefschrift, waaronder: Leo Brouwer, Jasper Carpaij, Paul Dixon, Hans Durinck en Mark Rogers.

Marin Kneyber, Martin, nu even voor jou. Bij aanvang van dit alles was ik wetenschappelijk naïef. Door met jou op één kamer te werken was dat snel over. Je enorme drive om onderzoek te bedenken, op te schrijven en te publiceren hebben mij in de juiste richting geduwd. Dank ook voor alle statistische adviezen daar waar het filter geen uitkomst bood. Hoofdstuk 6 en 7 zijn ontstaan uit een goede samenwerking.

Renata Sibarani-Ponsen, Renata wat moeten wij straks zonder jou? Jij hebt de gave werkelijk alles op de juiste manier te laten slapen. Dankzij jouw farmacologische ondersteuning en adviezen verliepen de experimenten succesvol. Je vermogen iedereen vertrouwen te geven is van al je goede eigenschappen, vind ik, de mooiste.

Hans van Vught, beste Hans, het onderwerp van dit proefschrift komt overeen met dat van je eigen proefschrift. Ik heb jouw versie met zeer veel plezier gelezen. Je enthousiasme voor mijn onderzoek en ook de kritische kanttekeningen erbij hebben mij een beter inzicht gegeven in dit onderwerp. Ik vind het dan ook een eer dat je het manuscript hebt willen beoordelen.

De overige leden van de promotiecommissie, prof.dr. A.P. Bos, prof.dr. J.B. van Goudoever, prof.dr. A.B.J. Groeneveld, prof.dr. S.A. Loer en prof.dr. D. Tibboel ben ik zeer erkentelijk voor het beoordelen van het manuscript.

Rob Peters en Roel Heslinga waren onmisbaar voor het aanpassen van software en het meedenken over de hardware, waarvoor veel dank.

Frans Plötz was initieel nauw betrokken bij het opzetten van de experimenten en becommentariëren van diverse manuscripten, waarvoor hartelijk dank.

Mijn collega’s Afke Robroch, Andra de Vries en Henny van der Meer hebben mij vaak net op tijd achter de computer vandaan gehaald. Jullie kijk op promoveren was vaak heel verhelderend.

Door de jaren heen is het gehele (verpleegkundig) team van de afdeling Intensive Care Kinderen onmisbaar geweest om dit alles tot een goed einde te brengen.

Page 150: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Dankwoord

149

Daarvoor ben ik alle dames, en enkele heren, dan ook heel veel dank verschuldigd! Nu dit alles is afgerond is er vast meer gelegenheid een aantal projecten op te pakken. Maar eens kijken wat de toekomst brengt.

Het lotgenoten contact met Mariëlle, balkon thuis, en Lukas, balkon ziekenhuis, gaf dit hele proces de nodige relativering. Bedenk beide ‘soms is wat onmogelijk lijkt alleen maar moeilijk’, en het komt goed, PhD at last.

Mijn paranimfen: Ellen en Petra. Ellen, jij maakte mij als noorderling ooit wegwijs op de multiculturele kinderafdeling van het Lucas. Petra, voor ons was dat andersom. Daaruit is een hechte vriendschap ontstaan. Jullie waren al paranimf nog voor er ook maar een letter van dit project op papier stond. Ik kan mij geen beter, getalenteerder, geestiger, krachtiger etc. duo voorstellen dat mij terzijde kan staan. Gaan we nu echt dansen?

Mijn ouders, zus, schoonouders en familie wil ik bedanken voor hun interesse, belangstelling en hulp met allerhande zaken. Mijn grootouders zouden trotser geweest zijn dan ikzelf.

Lieve Siem en Jet, met jullie is elke dag een feest! Ik denk niet dat jullie dit boek in één adem uit zullen lezen (om bij het onderwerp te blijven). Ik weet zeker dat het met die buurmannen op de voorkant wel op jullie eigen boekenplank terecht komt. Siem, rekenwonder, jij gaat mij later vast uitleggen wat er eigenlijk in dit boek staat. Jet, ‘maar dat kan ik eigenlijk al’, ook voor jou is er gelukkig nog genoeg te leren over. Welke aflevering van de buurmannen zullen we nu gaan kijken? Ineke, lief en tegenwicht. Je onvoorwaardelijke steun en geduld waren onmisbaar om dit proefschrift tot een goed einde te brengen. Maar ik ben je vooral dankbaar voor alles buiten dit boekje om. Komm gib mir deine Hand, we gaan iets leuks doen! Ich liebe dich!! A je to!

Page 151: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging
Page 152: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging

Curriculum vitae

151

Curriculum vitae

Marc van Heerde was born October 9, 1971 in Zwolle, the Netherlands. He graduated from the Zuyderzee College in Emmeloord in 1990. The same year he started his medical training at the University of Groningen in Groningen. After obtaining his medical degree in 1998 cum laude he worked as a medical resident at the department of pediatrics of the Sint Lucas Andreas hospital in Amsterdam (head dr. B.H.M. Wolf). In 1999 he started his training in pediatrics at the VU University Medical Center in Amsterdam (head prof. dr. J.J. Roord) and again at the Lucas Andreas hospital in Amsterdam (head dr. B.H.M. Wolf). Subsequently, in 2003, he acquired a fellowship in pediatric intensive care at the VU University Medical Center in Amsterdam (supervisors dr. D.G. Markhorst and dr. F.B Plötz). In 2006 he was officially registered as a pediatric intensivist and proceeded his career as a staff member at this unit. During his fellowship he became involved in the problem of spontaneous breathing during HFO ventilation. It led to a research project in close cooperation with the Faculty of Biomedical Engineering of the Czech Technical University in Prague, Kladno (supervisor dr. ing. K. Roubik) and the department of pediatric intensive care of the VU University Medical Center in Amsterdam (supervisor dr. D.G. Markhorst). Marc is married to Ineke Eckert. Together they have a son Siem (2004) and a daughter Jet (2006).

Page 153: Vrije Universiteit Amsterdam dissertation.pdfVRIJE UNIVERSTITEIT Preservation of spontaneous breathing during high-frequency oscillatory ventilation ACADEMISCH PROEFSCHRIFT ter verkrijging