We chat with Scott Weingart of Emcrit about the use of crash VA ECMO for the cardiac arrest patient.
Check out the REANIMATE course here!
Listen to the ED ECMO podcast on ECPR here
- ECPR candidacy may account for age, comorbidities, and code duration. Physiologic age is probably more important than chronological age. No-flow time without CPR should be very brief (witnessed is best), but low-flow time (with CPR) can actually be very long and still have good outcomes with ECPR. New systems should probably have stricter inclusion criteria, as numerous poor outcomes can endanger a fledgling program.
- The cause of arrest is usually not as important, partly because it’s often not known so early. ECPR can be a bridge to diagnosis and prognosis.
- One team should run the ACLS arrest while another handles the ECMO cannulation; it’s not possible to effectively do both. The cannulator should have their own ultrasound machine, and can function alone, although at least one skilled assistant is helpful. Mechanical CPR devices help by reducing energy in the room and reducing movement of the lower body; if not present, assign someone to manually stabilize the pelvis.
- Cannulation can be done by various services as long as they’re immediately available. Whoever it is should be comfortable using ultrasound. Cutdowns are probably not the preferred technique except in niche cases. A second service like CT surgery can arrive after a short delay to do the dilation and cannula placement if the in-department provider like EM or CCM can get initial access with smaller devices.
- Get ready by setting up equipment, position the ultrasound, and get sterile. As the patient arrives, have someone strip the clothes, expose the femoral region, and prep it, then get started with venous and arterial access.
- Vein vs artery cannot be distinguished without ultrasound, and can be difficult even with it. Don’t use anatomic location – use appearance. Arteries are thicker walled and small in cardiac arrest. TEE with a bicaval view to see your wire can be a huge help.
- The femoral artery should be accessed between the ligament and the bifurcation. Too high means RP bleeding risk; too low means potential for vessel damage. Similar for the venous access, although it’s more forgiving.
- Initially, place wires and then some kind of sheath, dilator, or line that will accept a larger, stiffer wire (Scott uses the Amplatz Superstiff). Going directly from needle to stiff wire is challenging and higher risk for vessel damage. This also means if you end up not proceeding to ECMO, you can just use the smaller sheaths for venous and arterial access.
- Even when a pulse returns, it’s often safer to proceed to ECMO in good candidates with a long arrest time. Supporting them through the next few days when they’re high risk of re-arrest, reperfusion injury, and other complications is likely to be safer than letting their heart do the work.
- Dilation for ECMO is similar to other dilation, just less forgiving. Follow the same consistent angle as the needlestick, constantly rack your wire, and consider dilating to a somewhat smaller cannula than in other VA ECMO situations, which is often tolerated post-arrest. Arterial cannulae of 17fr (women) to 19fr (men) or even smaller can achieve adequate flows, with venous cannulae of 19-23 Fr or even smaller.
- Goal: 5 minutes from first needlestick to active bypass.
- Ideally, one cannula per leg, but you can place both in the same side if needed. Certainly use the same side if using a cutdown.
- Venous cannula for the arrest patient should have the tip in the SVC (i.e. traversing the RA, not stopping before it). Use TEE to visualize this, or measure externally from groin to right nipple.
- Pumps can be pre-primed and sit waiting for 30-60 days in most cases; check manufacturer guidelines. Nurses can handle the pump with some extra training, at least for initial set-up, then transition care after 15 minutes or so to a perfusionist or ECMO-trained respiratory therapist.
- Pan-CT everyone. In fact, pan-CT all your cardiac arrests, as traumatic bleeding is common. Maybe do a coronary artery CT as well.
- Initial settings: 100% oxygen and titrate down quickly. Flow can be somewhat low compared to normal VA ECMO, allowing the native heart to keep some output and allowing smaller cannulas. Traditionally set sweep gas at roughly similar to bloodflow, but this tends to cause dramatic, rapid initial drops in PCO2, which may be harmful to a vulnerable brain; instead, start at a low sweep and gradually titrate it up.
- Do NOT prognosticate cardiac function early; recovery may happen late, and early withdrawal falsely affects your outcome figures from ECPR cases. The best numbers can only be achieved when the ECPR team continues to “own” the patient during their initial ICU course and doesn’t allow early withdrawal of ECMO.
- Neuroprognostication, conversely, tends to be easier; patients often stratify relatively early into clear good and bad outcomes. It should be established early on that families may want to pursue life support and that’s fine, but the team determines how long to continue ECMO, and it won’t be continued indefinitely.
- Economics: ECPR pump runs are short (<1 week usually), and reimbursement is all up front, so it actually pays well compared to many ECMO types, like long VV courses.
- The future: ideally, EMS would recognize good ECMO candidates and divert patients to ECPR centers. In rural areas, ED teams would be able to cannulate and start initially on ECMO, then transfer to larger referral centers.