Backflow perfusion of the distal limb in V-A ECMO

"Backflow" is the colloquial term referring to the perfusion of the limb distal to the insertion site of the return cannula in V-A ECMO, usually via a relatively thin 5-6Fr arterial cannula. It seems like an important aspect of ECMO to discuss, considering ECMO is half of "Mechanical supports: ECMO/IABP", an L1 topic of Section 2.1.4 in the second edition of the CICM Syllabus for the Second Part Examination. Or, rather, if one spends several hundred words on something like the water heater, then the backflow cannula should definitely get some attention. Confusingly, "backflow" is anterograde in terms of how the circulation is conventionally arranged, whereas the forward flow of the V-A ECMO circuit is retrograde, but this is the common terminology we seem to be stuck with, even though the ELSO people would prefer for us to use "minor flow" and "distal perfusion".

The only important material from below is:

For this niche topic, the Alfred ECMO distal perfusion cannula guideline is excellent as an easily readable quick reference guide, and for a peer-reviewed publication, one cannot go past Simons et al (2024). 

 How much blood does a leg need anyway

The cannulation of a large vessel (usually the femoral artery) to return arterial blood is an inevitable part of V-A ECMO. Unfortunately, arteries are rather smaller than veins. This creates an upper limit on the return cannula size, which is often the smallest available (eg. 17-19 Fr, 5.7-6.3 mm in diameter). The diameter of the common femoral artery in humans is often only 8-10mm, which leaves very little space for the perfusion of the limb. Consider: the remaining lumen (2mm?) will need to carry enough blood for the leg to remain healthy. 

How much would that have to be? This remains a guesstimate, on the basis of what we know has and has not worked historically. Ganz et al (1964) measured the femoral blood flows of adult Czechs at rest and determined it to be an average of around 560ml/min, ranging from 380 to 760. But none of these people were paralysed with muscle relaxant or dying of cardiogenic shock, and considering the usual femoral venous oxygen saturation is about 85% (Van der Schors et al, 2011), some might say the leg wasn't ever using all of that oxygen anyway. Certainly in practice it seems like people's limbs require less than that, with most authors aiming for at least 100ml/min and achieving often 150ml/min or more. 

The reader who has spent literally one second looking at the back of the packaging of a normal peripheral cannula will immediately recogise that this sort of flow rate should be achievable with a fairly thin device, perhaps as small as 18G (which suggests that a standard arterial line could be used for this purpose), but usually a 5 or 6 Fr arterial sheath ends up being the device of choice, mostly because the higher resistance of a narrow line ends up becoming the path of most resistance for the blood, and perfusion would be affected. 

The return cannula usually has a vented adaptor with a cap that takes standard 1/4 or 3mm Luer lock adaptors to connect the return line to the distal perfusion line. A flow measuring sensor is usually attached here to detect the flow of blood to the limb. The Alfred prefer thinner tubing here because the low flow rate means slow blood velocity, and that is a breeding ground for thrombi.

The optimum position for the cannula is as close to the ECMO cannula as possible, but also close to the bifurcation of the artery. The reason for this is that the area below the ECMO cannulation site, but above the backflow cannula site, is a weird liminal space where the trickle of anterograde flow coming down the partially obstructed femoral vein mingles with the opposing flow coming back up from the leg around the backflow cannula. What those flows will do is unpredictable, but the fact that they are flowing in opposite directions suggests that whatever ends up happening happens sluggishly turbulently, predisposing this region to thrombosis.

As see here, the rationale for cannulating at the site of the bifurcation (or even slightly distally) is that the area of stasis may potentially have some alternative flow pathway, i.e into the profunda femoris artery. The other benefit of this position is that it is much more difficult to accidentally cannulate the profunda femoris in this manner. However, despite whole minutes spent Googling this, no literature to support this practice had been found (though it appears to be well known).

The leg is on ECMO. What could go wrong?

Hmm.

• The flow is too high. One might think that having 300-400 ml/min or more is closer to the usual physiogical perfusion of this leg, but in fact this is probably not the case. One must remember that the ECMO cannula (and possibly also the nearby haematoma) is putting pressure on the venous egress from this limb, and venos congestion can develop. The resulting increase in compartment pressure can then lead to compartment syndrome and various other terrible things. The solution is to apply a flow reducing device of some sort to the flexible PVC tubing that supplies that cannula (a simple G-clamp is usually enough). If venous congestion does develop, Kasirajan et al (2002) proposes that we also cannulate the femoral vein in this limb, to allow venous drainage.
• The flow is too low. The clamp is fully released and still the limb is not getting anywhere near enough flow. There are several possible reasons for this, leaving aside the obvious "it's already thrombosed" or "you kinked the catheter" options. One which may be readily solved is the possibility that severe vasoconstriction has developed there, which can be relieved with some kind og vasodilator. A particularly effective cocktail is reported by Arroyo et al (2017); they used a mix of 5mg lignocaine, 0.5mg GTN and 2.5 mg verapamil (injected directly into the limb via the distal perfusion cannula).
• The limb is too large. The cannula may need to be upsized to allow enough flow, if the patient is much bigger than the usual range. Simons et al (2024) report occasionally needing to resort to a paediatric-sized ECMO return cannula (8-9 Fr) as a backflow cannula.  
• The limb obs are untrustworthy. The patient is sedated and paralysed, mottled from head to toe, and terminally cold. They cannot report paraesthesia or pain, cannot move their limbs actively, their pulses are impossible to palpate. Your options for objective monitoring now are serial Dopplers of pulses, serial limb circumference measurements, NIRS, and compartment pressure measurement using a pressure transducer and a 19G needle. A low-fi option could also be to put a saturation probe on the foot (the loss of signal, or decreased signal amplitude, could be an early warning of poor perfusion)
• The distal perfusion cannula is not functioning. Perhaps dislodged during a turn, or a dressing change, passive ROM physio, who cares (let's not get into the blaming and fingerpointing). The bottom line is, the limb is no longer perfused, and to re-access the femoral artery will be tricky. The options are:
  • Cutdown and re-access the femoral artery
  • Percutaneously cannulate the dorsalis pedis artery and then perfuse the limb retrograde
  • Resite the return cannula to the opposite leg (tough, if the opposite leg has a venous cannula in it already)