ECMO Essentials

In depth

The one-page summary above is what the PDF prints. These sections go further.

What the circuit does to the physiology

VV ECMO drains blood from the IVC or SVC, saturates it and returns it to the right atrium, where it mixes with uncaptured venous return. Arterial saturation depends on the captured fraction (ECMO flow ÷ native cardiac output), recirculation, and what the lungs add. At Hb 10 g/dL and a drainage saturation of 65%, each liter of saturated return adds about 45–50 mL of oxygen, so a resting VO₂ of 250 mL/min needs 4–6 L/min; ELSO General v1.4 sizes full adult support near 60 mL/kg/min.

CO₂ crosses the membrane about 20 times faster than oxygen, so its removal is set by sweep gas and membrane area rather than blood flow: sweep near 1:1 with blood flow gives normocapnia, and 1–1.5 L/min of blood flow clears most of the body's CO₂ (ECCO₂R). Oxygenation is capped by blood flow; once the membrane saturates the blood, more FdO₂ adds only dissolved oxygen, 0.003 mL/dL per mm Hg.

Peripheral VA drains the right atrium and returns blood retrograde up the descending aorta. Preload falls, so a distended RV and septal shift resolve, but LV afterload rises because the LV must eject into a circuit-pressurized aorta. Total systemic flow is ECMO flow plus native ejection, which the pulse pressure tracks. The two streams meet at a mixing zone: with a weak heart it sits in the ascending aorta and circuit blood perfuses the coronaries and brain; as the heart recovers it moves distally and the heart perfuses the arch vessels with whatever the lungs provide.

Circuit components and cannulation configurations

Flow through a cannula rises with the fourth power of its radius and falls with its length (ELSO General v1.4), so the long multistage drainage cannula is the bottleneck: adult VV drainage 21–25 Fr, return 19–23 Fr. Peripheral VA uses a 15–19 Fr arterial return cannula, the smallest that gives the needed flow since each size occludes more femoral artery, plus a 6–8 Fr distal perfusion cannula threaded antegrade into the superficial femoral artery.

Adult pumps are centrifugal: the impeller generates pressure and flow then depends on preload and afterload, so the same 3,500 RPM delivers 5 L/min to a full patient and 2 L/min to an empty one. The pump is non-occlusive: stopped on VA, arterial pressure drives blood backward through it, so a stopped VA pump is clamped at once.

The oxygenator is a polymethylpentene hollow-fiber membrane of about 1.8 m² in common adult units with an integrated heat exchanger. Its rated flow, about 7 L/min, is the flow at which 75%-saturated blood at Hb 12 g/dL leaves 95% saturated (ELSO Circuits 2022). The transmembrane pressure difference (ΔP) rises as fibers clot.

Patient selection and initiation

VV is for reversible lung failure, or a bridge to lung transplant, failing optimized protective ventilation. The usual thresholds are EOLIA's: P/F <50 for over 3 h, P/F <80 for over 6 h, or pH <7.25 with PaCO₂ ≥60 mm Hg for over 6 h; ELSO VV 2021 adds a Murray score ≥3 and asks that proning and paralysis be tried first when time allows. Relative contraindications: injurious ventilation for more than 7 days, uncontrolled bleeding, irreversible brain injury, no exit plan. The RESP score (Schmidt 2014) informs the decision without making it.

VA is for cardiogenic shock or arrest with a reversible cause or a transplant or durable VAD pathway: SCAI stage C–E shock with rising lactate on escalating support, cardiac index under 2.2 L/min/m², or refractory arrhythmia. Severe aortic regurgitation distends the LV under retrograde flow, a dissection can extend under the return jet, and severe peripheral arterial disease blocks femoral access. The SAVE score (Schmidt 2015) estimates survival; ELSO VA 2021 lists age, frailty, prolonged CPR and unrecoverable organ failure as reasons to decline.

Cannulation is percutaneous under ultrasound by Seldinger technique, with fluoroscopy or TEE to put the VV drainage tip near the IVC–RA junction and the return in the RA, or to aim a dual-lumen return port at the tricuspid valve. Heparin 50–100 units/kg is given unless bleeding forbids it (ELSO General v1.4). On VV, sweep starts near the blood flow on 100% O₂; on VA the blender is set for a post-oxygenator PO₂ near 150 mm Hg (ELSO VA 2021) and a right radial arterial line is placed.

Reading the console and the circuit

On VA the arterial tracing is the native heart's monitor: full support without ejection gives a nearly flat line at the ECMO-set MAP, and pulse pressure returns as the LV recovers or flow is reduced; ELSO VA 2021 reads a pulse pressure under 5–10 mm Hg as a closed aortic valve and distension risk. Pre-oxygenator saturation is an SvO₂ on VA (target above 66%) but is contaminated by recirculation on VV.

Troubleshooting problem by problem

ELSO's transport guideline defines the ECMO specialist as the person who changes pump settings; the bedside nurse recognizes the problem, protects the patient (clamps, ventilator, CPR readiness), calls and documents, and acts first for air, decannulation, a stopped pump or a ruptured line.

Recirculation fraction is (pre-oxygenator sat − true SvO₂) ÷ (post-oxygenator sat − true SvO₂) and needs a PA catheter, so at the bedside it is inferred: bright red drainage blood, pre-oxygenator saturation above 80%, and an SaO₂ that fails to rise as flow goes up. Causes are tips under about 10 cm apart, a rotated dual-lumen cannula, flow high relative to venous return, and hypovolemia; fixes are repositioning, lower flow, volume, or a second drainage cannula.

Complications: prevention, recognition, management

Bleeding affects roughly a third of adult runs (ELSO registry), driven by anticoagulation, platelet and fibrinogen consumption, acquired von Willebrand syndrome and the cannulation sites. ELSO 2021 targets UFH to anti-Xa 0.3–0.7 IU/mL or aPTT 1.5–2.5 × baseline (about 60–90 s); ACT correlates poorly on ECMO. Keep fibrinogen above 100–150 mg/dL and platelets above 50,000–80,000/µL (ELSO General v1.4 says 80,000), and transfuse at Hb 7 g/dL in stable VV (ELSO VV 2021). Bivalirudin 0.03–0.2 mg/kg/h or argatroban replaces heparin for HIT; both are titrated by aPTT and have no reversal agent.

For bleeding, hold or lower heparin (ELSO 2021: 4–6 h is reasonable, 12 h or longer for intracranial or surgical bleeding), replace fibrinogen and platelets, consider tranexamic acid, and get local control; coated circuits run without heparin for days while flow stays high. Intracranial hemorrhage occurs in about 2–8% of adult runs and kills about half; rapid PaCO₂ decline, thrombocytopenia, renal failure and hyperoxia are its risk factors (ELSO Neuro 2024).

Hemolysis is shear: a drainage pressure growing more negative, a pump-head clot, high RPM against an obstruction, or a return cannula too small for the flow. Plasma free hemoglobin above 50 mg/dL is significant (normal under 10), with rising LDH, falling haptoglobin, indirect hyperbilirubinemia, red-brown urine and rising potassium and creatinine (ELSO Circuits 2022; Bünger 2023). Inspect the pump head, lower RPM, improve drainage, and change the pump head if free Hb does not fall.

On VA an LV that cannot open the aortic valve against the retrograde pressure still fills from bronchial flow, Thebesian veins and any aortic regurgitation: LVEDP climbs, pulmonary edema and LV thrombus follow, and the subendocardium becomes ischemic. Signs: pulse pressure under 10 mm Hg, a closed aortic valve with spontaneous echo contrast, rising wedge pressure, pink frothy secretions. Management escalates from inotropes and the lowest tolerated flow to venting: IABP, Impella (ECPELLA), atrial septostomy, or a surgical LV or PA vent. Observational data favor venting (Russo 2019: 17 studies, 3,997 patients); randomized data are awaited.

Limb ischemia follows the femoral arterial cannula; ELSO VA 2021 places a distal perfusion cannula at cannulation (the ECPR guideline allows up to 4 h), with Doppler, NIRS and hourly limb checks afterward. Acute brain injury affects roughly 10–15% of adults on ECMO; ELSO Neuro 2024 asks for a baseline exam, EEG for unexplained encephalopathy, head CT for any new deficit, gradual PaCO₂ correction and PaO₂ kept under 300 mm Hg.

Daily nursing care by system

Lung rest lets the membrane carry gas exchange while the ventilator stops injuring. EOLIA reduced tidal volume to a plateau of 24 cm H₂O or less with PEEP ≥10, rate 10–30 and FiO₂ 0.3–0.5. ELSO VV 2021 suggests plateau under 25, driving pressure under 15, PEEP 10–15, rate 4–10 and FiO₂ under 0.5. Proning on ECMO did not shorten time to weaning in PRONECMO (Schmidt 2023: 170 patients, 44% weaned by day 60 in both arms).

• Neuro: pupils and exam every 1–4 h, daily sedation interruption when oxygenation allows, cerebral NIRS trended, CT for any change
• Cardiovascular: MAP ≥65 mm Hg on VA (ELSO VA 2021), under 80 after ECPR, on the lowest pressor dose; right radial line on femoral VA; daily echo for ejection, valve opening and cannula position
• Renal: fluid balance toward even or negative once stable; CRRT via separate access or the circuit (a clotting CRRT filter can falsely raise oxygenator pressures, ELSO Circuits 2022)
• Hematology: anti-Xa or aPTT every 6–12 h; CBC, fibrinogen, free Hb and LDH daily
• Circuit each shift: flashlight check of connectors, pump head and oxygenator inlet for fibrin (white) and clot (dark); ΔP trended; water in the gas outlet; clamps and hand crank at the bedside
• Cannulas: depth marks recorded each shift; Doppler pulses, color and NIRS hourly on VA; turns with one person on the cannulas
• Documentation hourly: flow, RPM, sweep, FdO₂, each pressure, circuit temperature, gases per protocol, every alarm and its response

Mobilization is a team event: a specialist on the circuit, a nurse on the lines, a therapist on the patient, and an abort plan. Dual-lumen jugular and axillary configurations allow walking; femoral cannulas are increasingly mobilized to sitting and standing with flow and pressures checked at each stage. Awake, extubated VV is the usual bridge to lung transplant.

Weaning and decannulation

VV readiness is rising compliance and tidal volume at unchanged pressures, a clearing film, and sweep and FdO₂ already reduced. The trial changes only the gas: blood flow stays (reducing it invites clot), sweep is turned off, and the ventilator is set to what the lungs must sustain (plateau 28–30 cm H₂O or less, FiO₂ 0.6 or less, PEEP 10–12 or less). ELSO VV 2021 describes a sweep-off trial of several hours; EOLIA required 6 h. Pass: PaO₂ above 60 mm Hg, acceptable PaCO₂ and pH, rate under 30–35 without distress.

Venous decannulation is done at the bedside: heparin often held 30–60 min, patient flat or head-down in an inspiratory hold or Valsalva so the jugular tract cannot entrain air, purse-string suture and 20–30 min of firm pressure. Cannula-site DVT is common, so most programs screen by duplex within the first week.

VA readiness is recovery on little pharmacologic support: pulse pressure above 10–20 mm Hg, MAP 60–65 or above, cleared lactate, recovering organs, no uncorrected valve lesion. Because flow is the support, the trial is a flow reduction in 0.5–1 L/min steps to about 1.5–2 L/min (ELSO VA 2021; lower flows clot unless anticoagulation is raised), with echo at each step and MAP, SvO₂ and lactate watched for 30–60 min at the lowest flow. Aissaoui 2011 criteria predict success: LVOT VTI ≥10 cm, LVEF ≥20–25% and lateral mitral S' ≥6 cm/s at minimal flow.

The arterial cannula comes out by surgical repair or a percutaneous closure device, and the limb is checked hourly for 24 h. Patients who cannot wean are bridged to durable LVAD or transplant, or support is withdrawn after a goals-of-care discussion. ELSO registry hospital survival is near 60% for respiratory VV, 45% for cardiac VA and 30% for ECPR.

Evidence and controversies

CESAR showed that referral to an ECMO center improves outcome, though only 76% of the referred arm received ECMO and the control arm had no protocolized ventilation. EOLIA missed its endpoint, but 28% of controls crossed over as rescue and 57% of them died; a Bayesian re-analysis (Goligher 2018) put the probability of benefit at 88–99%. ELSO VV 2021 adopted EOLIA's thresholds.

ARREST and Prague OHCA drove the AHA 2023 class 2a for ECPR in selected patients within a trained system. INCEPTION, with a median of about 74 min from arrest to flow across ten centers of varying experience, found no benefit, so the ECPR evidence reads as a test of system speed rather than of the device; ELSO ECPR 2021 asks for flow within 60 min.

In cardiogenic shock the randomized evidence is negative. ECLS-SHOCK found no mortality benefit with more than double the moderate-to-severe bleeding (23.4% vs 9.6%) and vascular complications (11.0% vs 3.8%); ECMO-CS agreed, and Zeymer's 2023 individual-patient meta-analysis (4 trials, 567 patients) found nothing. Critics cite 12.5% crossover, LV venting in only 5.8%, and a largely post-arrest population. DanGer Shock (2024) then showed Impella CP lowered 180-day mortality in STEMI shock (45.8% vs 58.5%), so practice has moved toward selective, vented support.

• Unsettled: routine versus selective LV unloading on VA (UNLOAD-ECMO is randomizing)
• Unsettled: low-intensity or no systemic anticoagulation on coated VV circuits
• Unsettled: oxygen targets on VA, temperature after ECPR (HYPO-ECMO 2022 was neutral), and when to neuroprognosticate (consensus waits at least 72 h)

Exam pearls

• VV is in series with the heart and supports only the lungs; a failing heart on VV needs inotropes or conversion to VA
• Blood flow sets oxygenation, sweep sets CO₂; FdO₂ is lowered on VA to avoid hyperoxia
• Femoral VA: right radial arterial line; low right-hand SpO₂ with pink legs is Harlequin syndrome
• VA pulse pressure under 10 mm Hg: the LV is not ejecting; think distension and venting
• Drainage chatter: lower RPM and give volume; never raise RPM into an empty vein
• Post-oxygenator PO₂ over 300 mm Hg on FdO₂ 1.0 is a healthy membrane; rising ΔP with falling flow is clot
• Anti-Xa 0.3–0.7 IU/mL; free Hb over 50 mg/dL is hemolysis; ECPR flow within 60 min; EOLIA 35% vs 46% (NS); ECLS-SHOCK no benefit