VV-ECMO weaning criteria in COVID ARDS with persistent RV dysfunction: when do you decannulate?

38M on VV-ECMO for 18 days following severe COVID ARDS (P/F ratio 52 at cannulation). Lung compliance has improved (from 12 to 28 mL/cmH2O). Current ventilator settings: FiO2 0.4, PEEP 10, TV 350 mL with P/F ratio 185 on minimal ECMO support (sweep 2 L/min, blood flow 3 L/min).

The problem: Bedside echo shows persistent RV dilation (RVEDD/LVEDD 1.1) with TAPSE 14mm. PA systolic pressure estimated at 52 mmHg. This has been unchanged for 5 days despite improving gas exchange.

Questions:

Do you mandate RV normalization before decannulation, or do you accept mild RV dysfunction if gas exchange is adequate?
What is your sweep gas trial protocol? We currently do 1-hour trials at sweep 0, but some centers advocate 4-6 hour trials.
Has anyone used inhaled iloprost as a bridge during the decannulation period to manage residual pulmonary hypertension?

This is a scenario we encounter with increasing frequency in our ECMO center. My approach:

1. RV dysfunction and decannulation: I do NOT mandate complete RV normalization. If gas exchange is adequate on sweep 0 for 4+ hours AND the patient tolerates it hemodynamically (no vasopressor escalation, stable CVP trend), I proceed with decannulation even with mild RV dilation. The rationale: once ECMO is removed and pulmonary blood flow normalizes, RV function often improves over days to weeks. The ELSO registry data supports this.

2. Sweep gas trial protocol: We use a staged approach:

Day 1: Sweep 1 L/min for 2 hours, monitor P/F, RV function on echo
Day 2: Sweep 0 L/min for 4 hours with continuous SpO2 and arterial blood gases at 0, 1, 2, and 4 hours
Day 3: If Day 2 was tolerated, repeat sweep 0 for 6 hours including a spontaneous breathing component
Decannulate on Day 4 if all trials passed

Key monitoring during trials:

Respiratory rate (must stay < 30)
P/F ratio must remain > 150 on FiO2 < 0.5
No new vasopressor requirement
Lactate stable

3. Inhaled iloprost: We have used it in 4 cases as a bridge. Started 24 hours before planned decannulation and continued for 48-72 hours post. The evidence is anecdotal but logical. It allows you to address residual pulmonary hypertension without systemic hemodynamic compromise.

One important pearl: check for ECMO-associated deep vein thrombosis before decannulation. We ultrasound both femoral and IJ veins. DVT at the cannula site occurs in up to 40% of patients and changes your anticoagulation planning post-decannulation.

Thank you Dr. Iyer. Sharing our La Paz protocol for comparison:

We use a combined cardiopulmonary weaning assessment that we published last year. The key difference from Dr. Iyer's approach: we incorporate RV strain biomarkers (serial NT-proBNP and troponin) alongside echocardiographic parameters.

Our decannulation criteria:

Sweep 0 tolerated for 6 hours
P/F > 200 on FiO2 < 0.5, PEEP < 12
TAPSE > 12mm (we accept values as low as 12, not the normal > 17)
NT-proBNP trending downward over 48 hours
No vasopressor escalation

On the RV question specifically: We analyzed our last 28 VV-ECMO decannulations. 9/28 had persistent mild RV dysfunction (RVEDD/LVEDD 0.9-1.2). All 9 survived to discharge, and 7/9 had normalized RV function by 30-day echo. This supports proceeding with decannulation despite mild RV dysfunction.

The one absolute contraindication to decannulation in our protocol: new or worsening RV dysfunction during sweep 0 trial (i.e., TAPSE dropping or new TR jet velocity increase). That suggests the RV cannot handle the full cardiac output without ECMO support.

2 Answers