Optimal PEEP

1
Optimal PEEP the final solution (Model-Based
Mechanical Ventilation for Intensive Care)
Geoffrey M Shaw1 J.Geoffrey Chase2 Chiew Yeong
Shiong2 Nor Salwa Damanhuri2 Erwin van
Drunen2 1Intensive Care, Christchurch Hospital,
New Zealand 2Mechanical Engineering, University
of Canterbury, New Zealand
2
Melbourne
Christchurch
3
Universities of Canterbury, Otago, and
Christchurch Hospital
4
Presentation Outline

• Acute Lung Injury (ALI) and Acute Respiratory
  Distress Syndrome (ARDS) in Intensive Care Unit
  (ICU)
• Treatment for ALI and ARDS (Mechanical
  Ventilation)
• Model-Based Mechanical Ventilation
• Minimal Model
• Elastance Model
• Results and Discussion
• Conclusion and Future Work

5
Acute Respiratory Distress Syndrome

• A syndrome of acute onset of respiratory failure
  with findings of bilateral infiltrates on chest
  radiograph, a partial pressure of arterial oxygen
  to fraction of inspired oxygen ratio (PaO2/FiO2)
  less than 300 (ARDS if less than 200mmHg) and the
  absence of elevated left heart filling pressure
  determined either diagnostically with a pulmonary
  artery catheter (pulmonary artery occlusion
  pressure of lt 18mmHg) or clinically (absences of
  evidence of left arterial hypertension)
• 1.3 to 22 per 100,000 (ALI 17.9 to 34 per
  100,000)
• Mortality is up to 70
• Ware et al. (2000). The acute respiratory
  distress syndrome
• Gattinoni, L. Pesenti, A. (2005). The concept
  of "baby
• lung
• Ferguson, N. D. et al(2005). Airway pressures,
  tidal
• volumes, and mortality in patients with acute
  respiratory
• distress syndrome.

6
Treatment for ALI/ ARDS
7
Model-based Mechanical ventilation

• Mechanical ventilation (MV) is the primary form
  of support for ALI/ARDS patients
• However, due to intra- and inter-
  patient-variability reduce the efficacy of
  general protocols
• Computer modelling can be used to identify and
  characterise patient-specific pulmonary mechanics
  and guide clinical decisions
• 2 Model-based Methods
• Minimal Model
• Lung Elastance Monitoring

8
Model-based Mechanical Ventilation Part 1

• A Minimal Model of Lung Mechanics

9
Minimal Model Recruitment

• Traditional Theory
• Isotropic Balloon like expansion followed by
  over-stretching
• Recruitment Theory
• Alveoli open or collapse
• Recruitment continues throughout the cycle
• Once recruited no significant volume change

10
Minimal Model Recruitment

• Alveoli do not behave like balloons

11
Threshold Opening Pressure/ Threshold Closing
Pressure

• Threshold Opening Pressure (TOP) Clinical
  Pressure when Alveoli Opens
• Threshold Closing Pressure (TCP) Clinical
  Pressure when Alveoli Collapse
• TOP gt TCP
• Crotti, et al. (2001). Recruitment and
  derecruitment during acute respiratory failure a
  clinical study

12
Minimal Model Development

• Based on the following Concepts
• Lung is modelled as a collection of lung units
• Either Recruited or Collapsed
• The state of every unit is governed by TOP and
  TCP
• The TOP and TCP for every lung unit assumed
  normally distributed.
• allowing fitting of a Gaussian Distribution
  Curve Mean and Standard Deviation

Original data sourced form
BERSTEN, A. D. 1998. Measurement of overinflation
by multiple linear regression analysis in
patients with acute lung injury. Eur Respir J,
12, 526-532.
13
Model Basics

• Mean Threshold Opening Pressure

• Standard Deviation of the distribution

14
Clinical Model Validation

• MEAN FITTING ERROR
• 1.62 - Inflation
• 4.42 - Deflation
• Capable of capturing patients fundamental lung
  mechanic
• Model TOP, TCP and SD

Original data sourced form
BERSTEN, A. D. 1998. Measurement of overinflation
by multiple linear regression analysis in
patients with acute lung injury. Eur Respir J,
12, 526-532.
15
Model - Application

• PEEP selection based on TOP and TCP concept
• TOP How much pressure required to open the Lung
  units
• TCP Maintain Recruitment
• Can this give us insight about the disease
  process?

16
Change in TOP or SD

• Monitor the Change of TOP or Standard Deviation
• Potential to group Patients based on TOP and SD
  information

17
Disease State Grouping (DSG)
How does the lung recover?
How does lung injury progress?

• A metric to classify patients disease state.
• Potential to guide MV treatment based on
  patients condition
• Theoretical Warrant investigation on TOP and SD
  relation with known patients disease state

18
Disease State Grouping (DSG)

• E.g. Bad cold

• vs

• Bird/Swine flu

19
Possible Examples in DSG
ARDS
H1N1
Healthy
Beginning of ALI
20
Example - Clinical

• Case Study 1
• 59y Male (survived)
• Pneumonia, COPD
• Day 0 (PEEP 12cmH2O)
• Auto PEEP 14cmH2O
• PaO2 114
• FiO2 0.4
• Average Mean TOP 45cmH2O
• Day 3 (PEEP 12cmH2O)
• Auto PEEP 8cmH2O
• PaO2 80
• FiO2 0.4
• No significant changes in Standard deviation -
  The lung state remains unchanged.
• Mean TOP drop with time The patients lung
  became less stiff compared to earlier.

SUNDARESAN, A., CHASE, J., SHAW, G., CHIEW, Y. S.
DESAIVE, T. 2011. Model-based optimal PEEP in
mechanically ventilated ARDS patients in the
Intensive Care Unit. BioMedical Engineering
OnLine, 10, 64.
21

• Case Study 2
• 69y Male (Deceased)
• Intra-abdominal sepsis
• Day 0 (PEEP 15cmH2O)
• Auto PEEP 11cmH2O
• PaO2 126
• FiO2 0.7
• Day 7 (PEEP 12.5cmH2O)
• Auto PEEP 2.3cmH2O
• PaO2 98
• FiO2 0.35
• Day 14 (PEEP 10cmH2O)
• Auto PEEP 1.6cmH2O
• PaO2 93
• FiO2 0.4
• TOP drops ? Lung is less stiff
• But SD increases meaning more lung (alveoli) are
  injured.

SUNDARESAN, A., CHASE, J., SHAW, G., CHIEW, Y. S.
DESAIVE, T. 2011. Model-based optimal PEEP in
mechanically ventilated ARDS patients in the
Intensive Care Unit. BioMedical Engineering
OnLine, 10, 64.
22
Model-based Mechanical Ventilation Part 2

• Continuously Monitoring Lung Elastance to Guide
  Mechanical Ventilation PEEP

23
Lung Elastance Monitoring

• Respiratory System Equation of Motion
• Paw Ers.V Rrs.Q P0
• Paw - Airway Pressure
• Ers - Respiratory Elastance
• V - Volume
• Rrs - Airway Resistance
• Q - Flow
• P0 - Offset Pressure (PEEP)
• BATES, J. H. T. 2009. Lung Mechanics An Inverse
  Modelling Approach, Cambridge University Press.

24

• What if... Respiratory System Elastance changes
  with Time during each volume increase?
• Paw (t) Edrs (t).V(t) Rrs.Q(t) P0
• Can we capture the lung condition with time?
• Continuous Monitoring of Lung Elastance/ Dynamic
  Lung Elastance and Resistance
• Integral Based Method (Similar to Multiple Linear
  regression)
• Monitoring the Elastance Trend may provide an
  opportunity to optimise PEEP
• SUAREZ-SIPMANN, F., BOHM, S. H., TUSMAN, G.,
  PESCH, T., THAMM, O., REISSMANN, H., RESKE, A.,
  MAGNUSSON, A. HEDENSTIERNA, G. 2007. Use of
  dynamic compliance for open lung positive
  end-expiratory pressure titration in an
  experimental study. Crit Care Med, 35, 214 - 221.
• CARVALHO, A., JANDRE, F., PINO, A., BOZZA, F.,
  SALLUH, J., RODRIGUES, R., ASCOLI, F.
  GIANNELLA-NETO, A. 2007. Positive end-expiratory
  pressure at minimal respiratory elastance
  represents the best compromise between mechanical
  stress and lung aeration in oleic acid induced
  lung injury. Critical Care, 11, R86.
• LAMBERMONT, B., GHUYSEN, A., JANSSEN, N.,
  MORIMONT, P., HARTSTEIN, G., GERARD, P. D'ORIO,
  V. 2008. Comparison of functional residual
  capacity and static compliance of the respiratory
  system during a positive end-expiratory pressure
  (PEEP) ramp procedure in an experimental model of
  acute respiratory distress syndrome. Critical
  Care, 12, R91.

25
Concept of Minimal Elastance

• During each breathing cycle, as PEEP rises,
  respiratory elastance (Ers) may fall as new lung
  volume is recruited faster than pressure can
  build up in the lung. This indicates
  recruitability
• If there is little or no recruitment, Ers rises
  with PEEP indicating that inspiratory pressure
  was unable to recruit significant new lung volume
  and now the pressure is, instead, beginning to
  stretch already recruited lung
• Hence, recruitment and potential lung injury can
  be balanced by selecting PEEP at minimum Ers
• Compared to a single, constant Ers value at each
  PEEP, identifying time-variant Edrs allows this
  change to be seen dynamically within each breath
  as pressure increases thus allowing a more
  detailed view of patients lung physiology.

26
Model-based Mechanical Ventilation Part 2

• Clinical Trials for Proof of concept

27
Clinical Protocol

• Patients underwent a protocol-based step-wise
  incremental PEEP recruitment manoeuvre (RM) using
  SIMV (Vt 500 ml) The ETT cuff pressure was
  inflated to 60 cmH2O to ensure there was no
  leakage so changes in FRC could be measured
• Baseline measurements were taken, then PEEP was
  decreased to ZEEP or reduced to a safe clinical
  level as determined by the PI)
• During the RM, PEEP was increased using 5 cmH2O
  steps until peak airway pressure reached at least
  45 cmH2O. Other settings were maintained
  throughout the RM.
• Each PEEP level was maintained for 1015 breaths
  until stabilisation before increasing to a higher
  PEEP level.

28
Patients Recruited in CHC Hospital

• A total of 10 patients have been included in the
  1st phase of the trial. (Still recruiting more)

Patients Sex Age (year) Clinical Diagnostic P/F Ratio (mmHg) FiO2
1 F 61 Peritonitis, COPD 209 0.35
2 M 22 Trauma 170 0.50
3 M 55 Aspiration 223 0.35
4 M 88 Pneumonia, COPD 165 0.40
5 M 59 Pneumonia, COPD, CHF 285 0.40
6 M 69 Intra-abdominal sepsis, MOF 280 0.35
7 M 56 Legionnaires 265 0.55
8 F 54 Aspiration 303 0.40
9 M 37 H1N1, COPD 193 0.40
10 M 56 Legionnaires, COPD 237 0.35
29
Patient 6 (Trauma)

• As PEEP Increases, Respiratory System Elastance
  drop until minimal before rising
• Minimal Elastance (Maximum Compliance) was
  observed at PEEP 15cmH2O
• The inflection line is identified as 510
  above minimal Elastance.
• Selecting PEEP at Minimum Elastance (Maximum
  Compliance) is not a new concept.
• Relatively few clinical trials have been carried
  out.

30
Example Variable PEEP with Respiratory System
Elastance
Pt 2 (Trauma) Minimal Elastance PEEP
15cmH2O Inflection PEEP 69cmH2O
Pt 6 (Intra-abdominal sepsis, CHF) Minimal
Elastance PEEP 15cmH2O Inflection PEEP
7.510cmH2O
Pt 8 (Aspiration) Minimal Elastance PEEP
25cmH2O Inflection PEEP 1218cmH2O
Pt 10 (Legionnaires, COPD) Minimal Elastance
PEEP 20cmH2O Inflection PEEP 1215cmH2O
31
Patient 6 (Intra-abdominal sepsis, CHF)

• Using Edrs with time, it is possible to identify
  the change of Respiratory Elastance within a
  breathing cycle
• A drop in Edrs will indicate the recruitment over
  pressure build up.
• An increase will suggest recruitment.
• The Respiratory system compliance within each
  breath can be monitored
• Edrs potentially provides higher resolution in
  monitoring the patients breathing condition
  compared to a single Elastance value within a
  breath

32
New Concept - Variable PEEP with Edrs
Pt 6 (Trauma)
Pt 2 (Trauma)
Pt 10 Legionnaires, COPD
Pt 8 (Aspiration)
33
Example Monitoring Edrs with time

• Measured Pressure (Blue Line)
• Model Pressure Fitting (Black Dots)
• Edrs within a breath (Red Line)
• What happens to Ers if PEEP Changes?

34
Comparing Lower and Higher PEEP in a Patient

• Ventilated at Lower PEEP
• Edrs within a breath drops, suggesting recruitment

• Ventilated at Higher PEEP
• Edrs within a breath increases, suggesting over
  distension

BERSTEN, A. D. 1998. Measurement of overinflation
by multiple linear regression analysis in
patients with acute lung injury. Eur Respir J,
12, 526-532.
35
Animal Trials in Belgium

• Animal Trials have been carried out to
  investigate the performance of the models.
• Healthy anesthetised piglet was ventilated with
  fixed tidal volume using Engström CareStation
  (Datex, General Electric, Finland).
• ARDS was induced using oleic acid.
• Subjects arterial blood gases were sampled to
  monitor the development of ARDS.
• Elastance (Ers, Edrs), and resistance (Rrs)
  using integral based method

36
Use of Electrical Impedance Tomography to compare
with our findings

• Electric Impedance Tomography (Collaborations)
• Zhao, Z., D. Steinmann, et al. (2010). "PEEP
  titration guided by ventilation homogeneity a
  feasibility study using electrical impedance
  tomography." Crit Care 14 R8.
• Minimal Model
• Sundaresan, A., T. Yuta, et al. (2009). "A
  minimal model of lung mechanics and model-based
  markers for optimizing ventilator treatment in
  ARDS patients." Computer Methods and Programs in
  Biomedicine 95(2) 166-180.
• Elastance Model
• Chiew, YS, Chase, JG, Shaw, GM, and Sundaresan, A
  and Desaive, T, Model-Based PEEP Optimization for
  Mechanically Ventilated ARDS Patients, BioMedical
  Engineering Online 2011.
• Cross comparison and validation

37
Animal trial results (unpublished)

• Blue Line - Integral Based Constant Ers.
• Black Line - Integral Based Median Edrs.
• Green Line - Multiple Linear Regression Median
  Edrs.
• Red Line - Ventilator Dynamic Elastance (?P/Vt).
• Pink Line - Ventilator Static Elastance ((Pend
  insp P0)/Vt) The ventilator has an
  inspiratory pause allowing estimation of Static
  Elastance.
• Blue, Black and Green Line Overlaps each other
• Computer Methods estimating Ers was able to
  reproduce the findings in ventilator.
• Change in Elastance was observed with the
  development of ARDS

38
Conclusion and Future Work

• The initial clinical trials indicate that the
  minimal model and respiratory elastance
  monitoring may be able to assist in the clinical
  decision for optimizing MV
• Minimal Model There is insufficient clinical
  data to determined the Disease Sate Groups. (What
  value is high TOP/SD?)
• Minimal Elastance Selection Proof of concept
  that warrants further investigation
• More trials to validate the performance of the
  model
• ARDS Animal model University of liege, Belgium
  (June 2012)
• Clinical trials open for recruitment (On going)
• Bedside and real time application (In progress)
• Tablet Software and Ventilator interface
  development

39
Research Collaborations?

• Main Collaboration on MV Research
• Cardiovascular Research Center, University of
  Liege, Liege, Belgium
• Intensive Care Unit, CHU Sart-Tilman, Liege,
  Belgium
• Institute for Technical Medicine, Furtwangen
  University, Germany
• Other Collaborations
• Intensive care and burn unit, University Hospital
  of Lausanne, Lausanne, Switzerland
• St-Luc University Hospital, Intensive care unit,
  Brussels, Belgium
• Intensive Care Unit, Clinique Notre Dame de
  Grâce, Gosselies, Belgium
• Intensive care unit, University Hospital of
  Geneva, Geneva, Switzerland
• Prospective Collaborations?

40
(No Transcript)
41
Thank you!
42
Supplementary material
43
Identifying Ers, Rrs and Edrs

• Multiple Linear Regression (MLR)
• Solving a Matrix
• Integral Based Method - Similar to MLR
• Instead of using data points of a curve, it uses
  the area under the curve
• More information and more robust to noise

• Paw Ers.V Rrs.Q P0
• We can identify Ers, and Rrs
• Using this Ers and Rrs from previous Equation
• Paw (t) Edrs (t).V(t) Rrs.Q(t) P0 can be
  solved.

44
Publications

• Sundaresan, A., T. Yuta, et al. (2009). "A
  minimal model of lung mechanics and model-based
  markers for optimizing ventilator treatment in
  ARDS patients." Computer Methods and Programs in
  Biomedicine 95(2) 166-180.
• Chiew, YS, Chase, JG, Shaw, GM, and Sundaresan, A
  and Desaive, T, Model-Based PEEP Optimization for
  Mechanically Ventilated ARDS Patients, BioMedical
  Engineering Online 2011.
• Sundaresan, A., J. Geoffrey Chase, et al. (2011).
  "Dynamic functional residual capacity can be
  estimated using a stress-strain approach."
  Computer Methods and Programs in Biomedicine
  101(2) 135-143.
• Sundaresan, A, Chase, JG, Shaw, GM, Chiew, YS and
  Desaive, T, Model-Based Optimal PEEP in
  Mechanically Ventilated ARDS Patients in the
  Intensive Care Unit, BioMedical Engineering
  Online 2011, 1064.
• Chiew, YS, Desaive, T, Lambermont, B, Janssen, N,
  Shaw, GM, Schranz, C, Moeller, K and Chase, JG
  (2012), Physiological relevance of a minimal
  model in Healthy Pigs Lung, 8th IFAC Symposium
  on Biological and Medical Systems, Budapest,
  Hungary. (In-Review) (Invited Paper)

45
Publications

• Mishra, A, Chiew, YS, Shaw, GM, and Chase, JG
  (2012), Model-Based Approach to Estimate dFRC in
  the ICU Using Measured Lung Dynamics, 8th IFAC
  Symposium on Biological and Medical Systems,
  Budapest, Hungary. (In-Review)
• Chiew, YS, Desaive, T, Lambermont, B, Janssen, N,
  Shaw GM and Chase, JG (2012), Performance of
  lung recruitment model in healthy anesthetized
  pigs, 2012 World Congress of Medical Physics and
  Biomedical Engineering, Beijing, China, May
  26-31, 1-page. (Accepted)
• Chiew, YS, Chase, JG, Shaw, GM and Desaive T
  (2012), Respiratory system elastance monitoring
  during PEEP titration, 32th International
  Symposium of Intensive Care and Emergency
  Medicine (ISICEM), Brussels, Belgium, March
  20-23, 1-page. (Poster Presentation)
• Sundaresan, A., Shaw, G. M., Chiew, Y.S. and
  Chase, J.G., PEEP in mechanically ventilated
  patients a clinical proof of concept,
  Australia-New Zealand Intensive Care Society
  (ANZICS) ASM, Taupo, New Zealand, March 31
  April 1, 1-page, (2011).