ARDS et Assistances respiratoires extracorporelles Matthieu Schmidt, MD, PhD Medical ICU iCAN, Institute of Cardiometabolism and Nutrition Hôpital Pitié-Salpêtrière, AP-HP, Paris Université Pierre et Marie Curie, Paris 6 matthieu.schmidt@aphp.fr

ECMO and ECCO2R? Same Technology Different Objectives…

Membrane lung O2/CO2 transfer ECMO for oxygenation ECCO2R for Decarboxylation CO2 transfer O2 transfer

ECMO and ECCO2R ECMO ECCO2R Large cannulas High extracorporeal flow >5000 ml/min Large membrane oxygenator Full blood oxygenation Full blood decarboxylation High technicity, ECMO center Double lumen catheter Low flow, respiratory dialysis 250-1000 ml/min Medium size oxygenator No blood oxygenation Partial blood decarboxylation Regular ICU

Full Flow ECMO Mid-Flow ECMO Low Flow ECCO2R

The evolving paradigm… ARDSnet strategy might not protect against tidal hyperinflation When Pplat remains >28-30 cm H2O Further decrease of Vt to reduce VILI From 6 to 5, 4 or 3 ml/kg IBW To decrease Pplat <25 cm H2O With sufficient PEEP to prevent lung derecruitment Resulting in a significant decrease in ∆P To decrease RR to <20… <15… <10????? Induced Hypercapnia controlled by ECCO2 removal “CO2 dialysis” Low-flow devices

VV-ECMO and ARDS

hospital mortality for severe ARDS 46%

Increasing Intensity of Intervention Mild ARDS Severe ARDS Moderate ARDS Increasing Intensity of Intervention 300 200 250 100 150 50 PaO2/FiO2 Low Tidal Volume Ventilation Higher PEEP Low-Moderate PEEP Prone Positioning NIV HFVO Inhaled NO ECMO Neuromuscular Blockade Lower Tidal Volume/Pplat The ARDS Definition Taskforce. JAMA 2012;307:2526-2533.

VV-ECMO To replace pulmonary function To allow the lungs to rest… To allow healing of the lungs…

2009… ECMO strikes back! Influenza A (H1N1)v09

17 25%)

17 25%) 14 (21)

17 25%) 6 (43)

Et al…

2013

ARDS…ECMO RCTs

ECMO provided only at the Glenfield Hospital, Leicester UK, 2001-2006 ECMO provided only at the Glenfield Hospital, Leicester Entry criteria: Adult patients (18-65 years) Severe, but potentially reversible ARDS Murray score ≥3.0, or Uncompensated hypercapnia: pH <7.20 Primary outcome measure Death or severe disability 6 months Analysis by intention to treat

Duration of MV > 7 days Lancet, 2009 Time from randomization to death Log rank p = 0.03 Exclusion criteria: Duration of MV > 7 days

no transfer on ECMO 6% died without ECMO

EOLIA objectives EOLIA trial designed to determine the effect of Early initiation of ECMO In patients with the most severe forms of ARDS

Inclusion Criteria American–European Consensus Conference definition for ARDS criteria Intubated and on MV for <7 days MV optimization before inclusion FiO2 ≥80% Vt = 6 ml/kg PBW Trial of PEEP ≥10 cm H2O

Inclusion Criteria One of the 3 following disease severity criteria PaO2:FiO2 <50 mmHg for >3 hours Despite potential use of inhaled NO, recruitment maneuvers Prone position, HFO ventilation, almitrine infusion PaO2:FiO2 <80 mmHg for >6 hours Despite similar criteria as above pH <7.25 with PaCO2 >60 mmHg for >6 hours Resulting from MV settings to keep Pplat ≤32 cm H2O Despite respiratory rate increased to 35/minute

Rescue ECMO for Controls Refractory hypoxemia SaO2 <80% for >6 hours Despite mandatory trial of Prone positioning AND Recruitment maneuver AND iNO or inhaled prostacyclin AND If the treating physician felt that Patient had no irreversible multi-organ failure AND ECMO might change the outcome

Primary Endpoint Relative Risk, 0.76, 95% Cl, 0.55-1.04; P=0.087 Hazard Ratio, 0.70; 95% CI, 0.47-1.04, P=0.074 by log-rank test

Key Secondary Endpoint Relative Risk, 0.62; 95% CI, 0.47-0.82; P<0.001 Hazard ratio, 0.48; 95% CI, 0.34-0.70, P <0.001 by log-rank test Percent D60 Treatment Failure Death in ECMO group patients; Death or Crossover to ECMO in control patients

Endpoint at D60 ECMO Group (N = 124) Control Group (N = 125)   Endpoint at D60 ECMO Group (N = 124) Control Group (N = 125) Median Difference (95% CI) Days alive and free of vasopressor use 49 [0–56] 40 [0–53] 9 (0 to 51) Days alive and free of cardiac failure (SOFA) 48 [0–56] 41 [0–53] 7 (0 to 51) Days alive and free of dialysis 50 [0–60] 32 [0–57] 18 (0 to 51) Days alive and free of renal failure (SOFA) 46 [0–60] 21 [0–56] 25 (6 to 53) Days alive and free of prone position 59 [0–59] 46 [0–57] 13 (5 to 59) Days alive and free of NO/prostacyclin 59 [0–60] 39 [0–58] 20 (4 to 59)

Crossover to ECMO in Controls Before crossover, of the 35 controls who had ECMO 9 had cardiac arrest 7 had severe right heart failure 11 developed renal failure requiring dialysis Venoarterial ECMO applied to 7 patients 6 under cardiopulmonary resuscitation

Control Crossover Outcomes

Rank-preserving structural failure time analysis - Controlling for crossover in controls Intention to Treat RFSFT Treatment Failure Hazard Ratio, 0.70; 95% CI, 0.47-1.04 P = 0.074 by log-rank test Hazard ratio, 0.51; 95% CI, 0.24-1.02 P = 0.055 by log-rank test Hazard ratio, 0.48; 95% CI, 0.34-0.70 P <0.001 by log-rank test

Which patient for VV-ECMO ?

Neuromuscular Blockade Mild ARDS Severe ARDS Moderate ARDS Increasing Intensity of Intervention 300 200 250 100 150 50 PaO2/FiO2 Low Tidal Volume Ventilation Higher PEEP Low-Moderate PEEP Prone Positioning NIV Inhaled NO ECMO Neuromuscular Blockade Lower Tidal Volume/Pplat The ARDS Definition Taskforce. JAMA 2012;307:2526-2533.

ECMO : potential indications EOLIA trial

Exclusion criteria Patients who are moribund and have any contra-indication to continuation of active treatment Irreversible ARDS Duration of high pressure and/or high FIO2 ventilation > 7d Intra-cranial bleeding Any other contra-indication to limited heparinisation

When to decide on the initiation of VV-ECMO for severe ARDS ?

not too late…. within the first 7 days on mechanical ventilation?

The Lancet Respiratory Medicine - July 2013

The Lancet Respiratory Medicine - July 2013

Predictive survival model (VV ECMO) Pappalardo et al, ICM 2013 Schmidt et al, ICM 2013 Roch et al, ICM 2013 Schmidt et al, AJRCCM 2014 Enger et al, Crit Care 2014

www.respscore.com

Potential role of such scores to help clinicians identifying “good” and “bad” candidates for ECMO: a simple bedside tool…

Potential role of such scores to help clinicians identifying “good” and “bad” candidates for ECMO NO

Potential role of such scores to help clinicians identifying “good” and “bad” candidates for ECMO NO YES

Potential role of such scores to help clinicians identifying “good” and “bad” candidates for ECMO NO ???? YES

We need high-volume centers for complex procedures…

Because high-volume ECMO centers have better outcomes…

with lower mortality was 22 (95% CI, 22–28) 0-19 20-49 >49 The minimum annual case load most significantly associated with lower mortality was 22 (95% CI, 22–28)

Where to perform VV-ECMO? Experienced centers in both ARDS and ECMO Better if cardiac and respiratory ECMO combined on the same site With experienced intensivists, cardiothoracic surgeons, perfusionists, nurses…. Minimum of 20 total ECMO runs per year ECMO programs should include a mobile ECMO referral team Available 24H/7D Nationwide or regional EMCO networks necessary

Conclusion VV-ECMO can rescue 50-80% of the patients with severe ARDS need to select patients who are more likely to survive ECMO should not be a futile prolongation of the ICU stay Important pathophysiological rational to use VV-ECMO early in severe ARDS in case of prone positioning failure < 7 days of MV potential role of predictive survival models