Notes from REANIMATE (part III): Pump management, ECMO critical care, and REBOA

More notes from REANIMATE. See Part I for patient selection, code flow, and pump setup; see Part II for cannulation.

Hooking to the pump

  1. Underwater seal on both cannulae. If bubbles, fix (small bubbles on venous side may be okay). Start oxygenator flow at 3 LPM. (Can use the 1:1 rule, i.e. sweep gas equals pump flow, but this often results in hyperventilation.)
  2. Release all clamps except a single clamp on the arterial side. This should be within reach of the pump operator (can reclamp if there are bubbles). Suggested not to place on the venous side, as sucking against complete occlusion causes fluid cavitation and pulls gas out of solution (the main problem of this is hemolysis).
  3. Start RPM at the “neutral” rate (varies by pump, often 1500 RPM). Slowly release the last clamp.
  4. Flow monitor should start to show flow. If not (or if retrograde flow is seen), increase RPM until you have forward flow. Check for bubbles, dark blood on venous side, and bright blood on arterial side.
  5. Increase RPM until flow is reasonable.
  • Blood is dark on both sides: oxygenator problem. Make sure you hooked up your gas.
  • Blood is bright on both sides: recirculation. Make sure recirculation bridge (if present) is closed. If so you likely have both cannulas in the same vessel, often the vein. Sorry.
  • Bubbles: look for the source. Run the circuit, ensure cannulae are in place (no holes are exterior). If problem resolved, can try to burp or aspirate the air out. If substantial, may simply need to replace the oxygenator or even the whole circuit. Big air volumes even on the venous side can eventually push through the oxygenator and reach the arterial circulation.
  • No flow: look for a clamp or kink. Make sure flow monitor is attached (with gel if needed). Check cannula position and patency (xray). If everything looks good, you may be clotted.
  • No increase in flow despite high RPMs: If flow maxes out early, your preload is a problem. There may or may not be chatter/shudder/chugging of the line here. You may also drop to zero flow (suction event), which is not always recognized, so you should set some low flow alarms. Your process should be:
    1. Reduce flow. Don’t leave it chattering, it causes significant trauma to the IVC at the suction holes.
    2. Trace out the circuit to check for kinks
    3. Check the external cannula depth (should be frequently measured by nursing so you can appreciate if it has shifted over time).
    4. Think about pneumothorax, tamponade, or abdominal compartment syndrome, all of which reduce preload.
    5. If none of the above, they need volume. But first thought should be whether they are bleeding, a common cause of hypovolemia in ECMO; check a Hgb! Then add volume, and come back up on flow when you can.

Pump management

initial post-pump care
  • Flows generally 4–5 LPM for full support (or 60 * kg). Increase RPM until the corresponding increase in flow begins to plateau; this suggests you are running out of preload. If you increase RPM and get no more flow, you are empty and should come back down. In ECPR, most speakers like to start with low flow, e.g. 2.5–3.5 and take some time to titrate up the sweep and flow to full support, due to worries about injury from rapid reperfusion. Patients are cold and sedated initially anyway.
  • Check a pre- and post-pump gas. The former is for SvO2 and determines perfusion; if poor, increase the flow. The latter is for pCO2; titrate the sweep to target a fairly normal range. Max sweep on the oxygenator is usually 15 LPM, and occasionally with large patients you might need more than this for ventilation — can use two oxygenators if so.
  • Post-pump gas can suggest the pO2, but to get a real idea, you need a right arm blood gas, which gives the true oxygenation to the aortic arch vessels (including carotids, coronaries, and right arm). Depending on heart/lung function (i.e. where the retrograde ECMO and anterograde cardiac flow mix), this may be much worse than the pump output. Get a right arm arterial line ASAP. An expert operator should do this, as it is important and not a good time to turn someone into a pincushion. Lots of different names for this out there: “differential hypoxemia,” “North-south syndrome,” “Harlequin syndrome,” etc.
  • Chest xray to check cannula position. Tip should be around the inferior cavoatrial juncture, or just before the hepatic vein. Some suggested going higher, i.e. tip in the RA or even SVC if you have a multistage catheter (the holes will drain the IVC and RA).
  • Urgent (not emergent, but not routine) echo to evaluate cardiac function, tamponade (iatrogenic perforation), and presence of aortic regurgitation.
  • Once the heart has perfused for a bit, shock out any arrhythmias. Often they break easily. Sometimes they need antiarrhythmics.
  • Get a distal perfusion catheter placed in the cannulated leg. Generally IR should do this. Can use an anterograde cannula (7 French or so) in the ipsilateral SFA, or can do a retrograde catheter in the dorsalis pedis, which may have a bit less bleeding risk. 20% of patients will develop leg ischemia, and they do very poorly; “if they leg dies, the patient dies.” Don’t watch and wait, just place a catheter in everyone.
  • The pump is afterload sensitive, so if you have severe post-ROSC hypertension due to epi, use nicardipine or similar to manage it.
  • Anticoagulation: most start with a bolus of 5000u heparin (can do it weight-based, but people seem to prefer simplicity), then a drip. Other agents could be used. Interestingly, Dr. Bartlett is considering moving toward DOACs. Surely reversal would be tricky?
    • ACT target: 180–200
    • PTT target: 45–55
Managing intrinsic cardiac function
  • Initially total cardiac rest is probably good to minimize myocardial oxygen demand. Increase pump output until the art line is flat, ETCO2 shows nothing (no blood reaching the lungs!), or echo shows no output.
  • Once things cool down, the heart may begin to recover, which is good. Zero systolic output results in intracardiac blood stagnation which causes thrombi (even when anticoagulated). More importantly, the LV distends, eventually leading to horrible hemorrhagic pulmonary edema (gallons of fluid pouring out from the lungs) and usually death; many speakers emphasized the morbidity of this. Dr. Bartlett likes to track this using the Swann to monitor PA pressures. Trending CVPs might be a surrogate.
  • Several speakers suggested a goal of one aortic opening every 3–4 electrical systoles. This should empty the LV effectively. Hitting this goal requires adjustment of flow (allowing the heart some preload) and usually some inotropy (drug of choice). If you can’t empty the heart, Impellas are becoming popular, or you can vent the LV (and Y the drain into your venous return), or even do atrial septostomy.
  • Problem: heart recovering before lungs. Native CO through severely shunted lungs causes hypoxemic blood ejecting into the arch and malperfusing the brain and coronaries. Anyone with intrinsic cardiac output needs close vent monitoring and right arm ABGs. Can follow right arm SpO2 but less reliable. If the heart is too active, can try to slow it (esmolol?). Can add an arterial cannula to the right IJ or subclavian and Y it into the VA circuit (“VAV” or “VVA”) to perfuse the arch. Can also just switch to VV if the heart is doing well.

Critical care on ECMO

Address underlying causes of arrest
  • Most arrests should be cathed if there are signs of ischemia, or cryptogenic cause.
  • Suspected PE: usually the best thing to do is nothing. The patient is fully anticoagulated, the RV is unloaded, and their cardiac output is supported. Confirmatory CTA is tricky (the pump sucks up the contrast and usually needs to be paused) and adds little value. Just wait.
General patient management
  • Post-arrest patients will obviously have metabolic acidosis. Weingart’s answer is to give bicarb (slow pushes); unlike in other cases, here you can increase the sweep to clear the resulting CO2. However, now that perfusion has normalized, lactate should quickly clear and acidosis should resolve, so don’t overshoot.
  • During and after cannulation, need great caution about any open ports on a central line. Nothing should ever be left open to atmosphere, as the highly negative venous pressures on the pump can rapidly entrain air. Any drips should be on pumps; freeflow fluid bags can be sucked dry including the bubble.
  • Get a central line; you can technically give meds through the ECMO circuit but should not make a habit of it. Right IJ is probably best. Use serious air embolism precautions, including steep Trendelenburg. Subclavian is okay for experts, but beware bleeding. Femoral is highest risk for air embolism; in that case perhaps reduce flow during placement. Maybe use the Raulerson syringe?
  • If available, you can place perfusion monitors on the head, right arm, and cannulated leg. Trending them is… somewhat useful.
  • Watch out constantly for bleeding, which is ubiquitous. Also trend a daily LDH for hemolysis. Often this is due to a catheter issue (e.g. a small kink) and is correctable.
  • Despite the frequently less-than-sterile placement in ECPR, line infections at the cannulas seem infrequent. Dr. Hackmann does sometimes give Ancef if it was very dirty.
  • If pneumothorax develops (e.g. after line placement), try to avoid messing with it. Usually you can watch with serial CXR. If it causes tension, it does need to be drained, but try a small pigtail. If hard tube is needed, consider placement in the OR. Exsanguination from chest tube placement is not uncommon.
  • Ambulation on ECMO is okay. Often they are too sick on VA, VV is more common. All-IJ is best, but femoral lines are still doable; they cannot flex their hip, however. If unable to ambulate you can do some tilt-table.
  • New extremes in circuit pressures and drops in flow may be due to clots. Most common in axillary cannulas due to the angle they make. Pumps can clot as well.
  • CRRT can and probably should be linked to the ECMO circuit, although technically most of the vendors do not technically recommend it. Various ways to do this; some places actually bridge over the oxygenator with the CRRT circuit. Be cautious, as open ports are common on the CRRT circuit and can be a source of air.
  • Pressors: target a MAP, as usual. Consider a slightly higher goal, such as 70–80 in post-arrests. Can alternately calculate an SVR (goal 800–900), using (MAP/flow) * 80, but not sure if it adds much.
  • If getting inadequate flows, you can add an additional venous or arterial cannula to the circuit. Place the line into the contralateral fem or the right IJ, clamp your cannulas and turn on the recirculation bridge, then Y in the new line.
  • Transfusion: Controversial. Arguably a Hgb of 7–8 is fine. You must balance oxygen carrying capacity versus viscosity. Some speakers prefer higher goals, at least >10; Bob Bartlett likes 15. Remember that you do need good preload, and obviously crystalloids have a volume of distribution that tends to make you chase your tail.

The heater/cooler is quite heavy and patients can sometimes be transported without it. Just beware as they will rapidly cool down.

Technically, patients on ECMO may not need a ventilator. If it’s creating a big transport challenge you may simply be able to clamp the ET tube.

Ventilator management

Goal is purely recruitment, and enough oxygenation to avoid differential hypoxemia. High PEEP, low volumes, high FiO2.

Simple approach: “10s all around.”

  • PC 10 (or use about 4cc/kg PBW)
  • PEEP 10
  • RR 10
  • 100% FiO2

Getting off ECMO

Once things look good, do a weaning trial. Turn down flow by whole numbers (4.5 -> 4 -> 3 -> 2), doing a serial echo at each stage. Don’t go below your minimum (1500 RPM or whatever your pump requires), and don’t stay longer than 5-10 mins at those low speeds. If native function is good, stop the pump and pull your lines.

Probably get some dopplers of the legs to look for clots afterwards.

Closing the arteriotomy
  • Pressure alone usually okay for non-huge venous lines. Hold for about 3x the French size (e.g. 21 minutes for 7 Fr) as a starting rule. Pressure is trickier with poor site selection (too high or low); you need the femoral head to press against
  • Hemostatic patches like FemoStop or percutaneous closure devices like Mynx, Angioseal, or Starclose are an option for <8 French arteriotomies.
  • Suture closure devices like Proglide are okay up to 10 French. Best if placed at the time of cannulation (although you can get it in there later); you can just park the stay sutures on the skin until ready to close later.
  • Open surgical repair: always an option and mandatory for big holes. Transverse closure is preferred as it won’t cause any stenosis, but you may need to do longitudinal if an endarterectomy is needed (e.g. if there’s too much atheresclerosis and you need to make space for closure).


Generally need larger cannulas, as patients often have quite hyperdynamic flow (e.g. in sepsis) and you need to keep up. Ventilation is typically the bigger challenge. You may need additional cannulas for adequate flow, particularly inflow.

Usually oxygenation should be quite easy. If there is hypoxia, check a pre- and post-oxygenator gas; if the latter is low, it’s probably an equipment issue.

Can you do elective VV ECMO for difficult airways? Maybe…

VV Cannula tips

For a right IJ, use the same depth as a regular triple-lumen. Use ultrasound to avoid placing a big cannula through the SCM, can cause bleeding. Goal is to place tip at caudal cavo-atrial junction.

For dual femoral placement, you can’t use multi-stage catheters, and you need to get the lengths correct, or you’ll get lots of reflow. Place both wires first and then cannulae.

Many attendees have stopped using the dual-lumen Avalon catheter due to various issues.


New ER-REBOA catheter fits into a 7 Fr sheath, unlike old 10 Fr that required vascular repair and placement over wire.

Seem to be similar outcomes to thoracotomy + crossclamping, but obviously a less morbid intervention.

AORTA registry suggests best results seem to be when used just pre-arrest (not after arrest, not super electively).

Probably okay to leave balloon inflated <100 minutes. <60 would be less controversial, but either way you can probably go longer than crossclamping. Maybe you can do longer in Zone 3? Unclear. Either way, it’s clearly a bridge to reaching the OR and surgeon presence, not a bridge to transfer or longer delays; ischemia and reperfusion kills you then.

The company does not recommend partial occlusion, but Dubose thinks it’s okay (maybe it’s because they’re releasing a new device for this purpose soon). He transduces the side port of the sheath so he can compare pressure above/below the balloon and inflates until bleeding controlled.

Remember that in OB bleeding from an ovarian ectopic, perfusion is likely from higher than you think, so do a Zone 1, not Zone 3.

  1. Place a femoral A-line into the CFA early, while the vessel isn’t too flat.
  2. If decision to use REBOA, rewire it to a 7 Fr sheath.
  3. Decide on zone 1 (balloon sitting atop diaphragm), best for abdominal or unlocalized bleeding, or zone 3 (balloon sitting above iliac bifurcation), best for known pelvic bleeding (e.g. pelvic fx with negative FAST). Zone 2 in between not recommended as this is usually the most atherosclerotic territory.
  4. Measure catheter from sheath to: sternal notch (zone 1) or xiphoid (zone 3). Measure both so you can change zones later if needed.
  5. Insert cannula, using orange sheath to open valve. Once balloon is through sheath you can push back the sheath (can peel it off, but no need, and better to hold onto it just in case.) Place to target depth. Transduce pressure from the catheter tip.
  6. Inflate balloon. Dubose’s method: mix saline and IV contrast and draw up 24cc (the max the balloon can hold; never push more than this). Use a large syringe (30cc) and use the same size each time so you can get to know a consistent feel. Inflate until you feel increased resistance. Should also see an increase in the aortic pressure.
  7. Hold in place at the skin, or the pressure wave will push out the balloon. Secure with a friction clip (like a triple lumen). Xray to confirm placement. The ends of the balloon will flatten out (losing your “olive” shape) if overinflated.
  8. Removal: remember to withdraw all volume from the balloon before pulling. (Useful tip: syringes will hold suction on their own if you withdraw the plunger fully.) Drop the volume slowly at first, however; use a small syringe and pull fractions of a cc at a time. Even 1 ml is a lot in these patients due to reperfusion.
  9. Pull and hold pressure; no need for a vascular repair.


Big summary points

  • People are really afraid of LV dilation; ECMO fixes most problems but not this. Watch for it closely and manage early.
  • Bleeding is always a concern. Watch for it.