Model-based elastance and optimal peep selection

1
Model-based elastance and optimal peep selection

• Geoffrey M Shaw1
• Yeong Shiong Chiew2
• J Geoffrey Chase2
• Ashwath Sundaresan2
• Thomas Desaive3
• 1 Dept of Intensive Care, Christchurch Hospital
• 2 Dept of Mechanical Engineering, University of
  Canterbury
• 3 Thermodynamics of irreversible processes,
  Institute of Physics, University of Liege,
  Belgium

NZ ANZICS Dunedin March 15 2013
2
Declarations

• Member, Medical Advisory Board, Baxter NSW
• Director, Lifevent Medical
• Hold world-wide patents on Low flow CPAP device
• Multi-modal high frequency ventilator

3

• What do marriage and ARDS have in common?...
• You only have a 50 change of getting out of
  either alive!

4
Acute lung injury /respiratory distress syndrome

• 10 to 80 cases per 100,000 persons per year
• Mortality rate for ARDS 30 to as much as 80
• In the US
• 190,600 cases of ALI, including ARDS, per year
• 74,500 associated deaths per year
• 3.6 million hospital hours per year
• Thus, 12000 in the ANZ ICUs with 4800
  associated deaths

http//www.meddean.luc.edu/lumen/MedEd/medicine/pu
lmonar/diseases/pdis3.htm
5
Acute lung injury /respiratory distress syndrome

• Inflammation Capillary leak Alveolar collapse

http//medicinembbs.blogspot.co.nz/2011/02/obstruc
tive-vs-restrictive-lung.html
6
Superimposed pressure
(modified from Gattinoni)
7
Acute lung injury /respiratory distress syndrome
8
Current practice based on one size fits all...
High Vt
9
Current practice based on one size fits all...
10
The conundrum......
Too high Injure healthy lung tissue
PEEP
Too low No lung recruitment
11
Aim...

• To develop a model-based solution to guide PEEP
  selection in mechanical ventilation that
• Predicts patient-specific response to treatment
• Balances risks of overstretch vs. derecruitment
• Monitors disease state
• Optimises work of breathing

12
Study design

• Ten patients with ALI/ARDS (PaO2/FiO2 150-300)
• Protocolised recruitment manoeuvre PEEP
  increased in 5 cmH2O increments until Paw 45
  cmH2O

PEEP cmH2O
30
25
20
15
10
5
Time
10 breaths
13
Study design

• Participant is sedated and paralysed for duration
  of RM
• PB 840 ventilator Vt 400-600 ml
• Pnuemotachometer to measure pressure flow
• (Hamilton Medical, Switzerland flow sensor)
• National Instruments USBB6009 and Labview Signal
  Express to obtain measurements at 100Hz
• (National Instruments, TX, USA)
• Analysis performed using MATLAB
• (The Mathworks, Natick, Mass, USA)
• Laptop (Dell)

14
Model-based analysis

• Validated relevant recruitment model using single
  compartment
• Captures patient-specific fundamental lung
  mechanics in real-time to identify
• Constant lung elastance (E lung)
• Dynamic lung elastance (Edrs )
• Compliance 1/Elastance

15

• Some scary maths on next side, look away if you
  wish

16
Model-based analysis
Lung Component
Airway Component
R

• Paw Elung V Rlung Q PEEP

Paw (t) Edrs (t)V (t) Rlung Q (t)
PEEP
17
Model-based PEEP selection

• During each breath Elung will fall if new lung
  volume is recruited faster than the pressure
  builds up ? RECRUITMENT
• If little or no recruitment occurs Elung rises
  with PEEP, because at that pressure level there
  is no further recruitment recruited lung is now
  beginning to stretch
• Hence recruitment and potential lung injury can
  be balanced by selecting PEEP at minimal Elung
• Edrs allows this change to be seen within a
  breath providing a more detailed view.

18
Three approaches to PEEP selection

• Minimum Elung and Edrs
• Over all peep levels (and pressure for Edrs)
• Minimum Edrs Area.
• Integrating Edrs over each breath for each PEEP
  level is more clinically relevant and is
  proportional to WOB
• Inflection Method
• The PEEP that corresponds to the point where
  Edrs or E lung is 105-110 of the minimum as
  the point of inflection where there are
  diminishing returns

19
Patients
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
20
Edrs
APE Absolute Percent Fitting error
Similar results for Elung (APE 5.9 ) and Edrs
Area
21
Dynamic lung elastance Edrs vs. PEEP
Pt 6 (Abdominal sepsis, CHF)
Pt 2 (Trauma)
Pt 10 (Legionnaires, COPD0
Pt 8 (Aspiration)
22
Constant lung elastance Elung vs. PEEP
Pt 2 (Trauma) Minimal Elastance PEEP
15cmH2O Inflection PEEP 69cmH2O
Pt 6 (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
Similar results for Edrs Area
23
Response to PEEP in H1N1
24
Low heterogeneity recruitment
PEEP (cmH2O) 5 10 15
Edrs (cmH2O/L) Median IQR 40.5 36.4-52.8 39.9 35.8-48.7 31.2 30.2-33.6
E drsArea (cmH2Os/L) 55.2 51.3 38.3
Elung (cmH2O/L) 39.1 38.2 31.1

• Reduced Edrs range with increased PEEP shows
  improved heterogeneity

25
Clinical vs. modelled Edrs Area (inflection) PEEP
Modelled PEEP
Clinical PEEP
Patient No.
PEEP (cmH2O)

• In all but one patient (pt 2), clinically-selected
  PEEP was significantly less than a model-based
  estimate using minimal elastance

26
Limitations

• Uses a single compartment model, so results are
  for the whole lung. Thus regional differences
  are not detected
• Patients required sedation/paralysis. Model
  assumes pleural pressure 0, which may not
  always be valid
• Elung , Edrs are effective respiratory system
  elastance values within these metrics are
  unmeasurables such as regional resistances
  within the lung
• Requires validation in prospective clinical
  trials

27
Summary (1)

• This model-based approach provides
  patient-specific insight not easily directly
  measurable
• Method can be easily implemented with minimal
  PEEP titrations. Currently PEEP is selected on
  descending limb of RM
• Obviates the need to use a one-size-fits-all
  maximal recruitment technique, typically with
  pressures up to 55cmH2O
• Modest changes in PEEP will detect Edrs changes
• Can be used to trend changes in lung condition

28
Summary (2)

• Current methods of PEEP selection poorly agree
  with model-based results. This suggests nearly
  all patients are sub-optimally ventilated, which
  may significantly impact on outcomes
• Results from large RCTs to date have not
  optimised PEEP to minimal elastance and therefore
  we may need to re-think the interpretation of
  these findings- 6 ml/kg cannot be optimal for all
  patients

29
Where are we going? What does the future hold?
30
(No Transcript)
31
Elastance changes following recruitment
Successful recruitment manoeuvre
Two unsuccessful recruitment manoeuvres
Elastance (cmH2O/L)
PEEP (cmH2O)
Time (mins)
Time (mins)
32
Elastance trending
Improving condition
Worsening condition
Elastance decrease
Elastance (cmH2O/L)
PEEP (cmH2O)
Elastance (cmH2O/L)
Elastance increase
Time (mins)
Time (mins)

• Decreases in elastance over time means patient
  condition is improving
• Good PEEP selection
• Increases in elastance over time indicates
  patient is derecruiting
• PEEP change or recruitment maneuver may be needed

33
dFRC Model Monitoring patient lung volume
1600 1400 1200 1000 800 600 400 200 0
dFRC
Recruited lung
Derecruited lung
0 5 10 15 20 25 30
Airway pressure (cm H2O)
Picture (modified) sourced from UCSMT
34
dFRC Model Monitoring patient lung volume
Patient condition constant, average dFRC constant
Volume (L)
Pressure cmH2O
Patient condition improving with more lung
availability
Time (h)
35
Clinical interface
Edrs following RM
36
Acknowledgements

• Yeong Shiong Chiew
• J Geoffrey Chase
• Ashwath Sundaresan
• Thomas Desaive
• Richard Fernando
• Laura Badcock
• Sarah Poole
• James Williams

37
Acknowledgements

• Intensive Care Staff Christchurch Hospital