Chapter question: There is a large cannula in this patient's femoral artery and they are sedated. How do I know the leg is alive, and what do I do the moment it isn't?
Evidence search date: 10 September 2026.
Two canonical reviews were retrieved in full this session and carry most of the numbers below: the 2019 Critical Care narrative review of limb ischaemia in peripheral VA ECMO (DOI 10.1186/s13054-019-2541-3), and the 2024 Perfusion review of the evolution of distal limb perfusion management (DOI 10.1177/02676591241236650). Around them sit five meta-analyses and eight cohort studies identified through a structured search.
Primary sources: ELSO Red Book 6th edition Ch 26 and Ch 33; ISCCM Manual Ch 41; Taha Ch 7; ECPR and Resuscitative ECMO Ch 5.
This chapter's central finding is uncomfortable. Every meta-analysis of observational data says a prophylactic distal perfusion cannula prevents limb ischaemia. The largest and longest single-centre series say it does not — and they identify a different dominant risk factor altogether. §17.6 sets both out.
What this chapter covers — and what it does not
This chapter owns | Deferred to |
All six mechanisms by which a limb dies on ECMO, arterial and venous; incidence and the definition problem; limb ischaemia as a mortality marker; the risk-factor set and its surprises; surveillance — why clinical examination and Doppler both fail, and what NIRS adds, with thresholds; staging a threatened limb; distal perfusion technique, sizes, routes and timing; the prophylactic-versus-selective controversy; venous congestion; the conservative and invasive treatment ladders; upper-limb hyperperfusion; dissection, pseudoaneurysm, retroperitoneal haemorrhage and late arterial stenosis; two devices in one limb | Choosing the arterial site and sizing the cannula → Chapter 12
The 4-hour distal perfusion deadline → Chapter 14
Micro-axial pump management → Chapter 80
Cannula malposition, kinking and migration → Chapter 34
Circuit thrombosis → Chapter 31
Systemic bleeding and its management → Chapters 36 and 45
Anticoagulation targets and monitoring → Chapters 39–42
Decannulation and vessel repair → Chapter 18
Mobilising a patient with a distal perfusion cannula → Chapter 59 |
17.1 Six ways a leg dies, and only two of them are the cannula
The standard mental model is mechanical: a big tube in a small artery obstructs flow. That is true and it is incomplete, and the incompleteness is why prophylactic hardware alone does not solve the problem.
Physiology — the six mechanisms, and which interventions touch which
A 2024 narrative review sets out the mechanisms of limb ischaemia on ECMO as: arterial obstruction; cannulation injury; loss of pulsatile flow; thromboembolism; venous stasis from compressive obstruction by large venous cannulas; and systemic vasoconstriction from shock and from pharmacological vasoconstrictors.
Sort them by what fixes them and the shape of the problem changes:
Mechanical, arterial (1, 2) — the cannula occupies the lumen, or damaged it going in. A distal perfusion cannula addresses this, and cannula size and technique prevent it (Chapter 12).
Mechanical, venous (5) — the drainage cannula obstructs outflow. A distal perfusion cannula does nothing for this, and may make it worse by adding inflow to a limb that cannot drain (§17.7).
Haemodynamic (3, 6) — non-pulsatile flow and vasoconstriction. No cannula fixes these. They are treated by turning vasopressors down, and Chapter 13 has already established the lever that makes that possible on VA ECMO — flow is the vasopressor.
Thrombotic (4) — embolus or in-situ thrombosis in a low-flow segment. Treated by anticoagulation, and, on the emerging evidence in §17.4, possibly by antiplatelet therapy.
Only two of six mechanisms are addressed by the intervention that dominates the literature. Hold that thought through §17.6.
Clinical pearl — the limb question differs at each end of the patient, and now at each end of the vessel
Chapter 12 established the first axis: femoral arterial return risks ischaemia below; axillary or subclavian return risks hyperperfusion above (§17.9). Chapter 16 added a reason to reach for the subclavian route, which means the trade is now made more often.
This chapter adds the second axis, within one leg: the artery can be obstructed, or the vein can be, and they present almost identically — a swollen, dusky, cool leg. The treatments are opposite. Putting more arterial inflow into a congested limb is the specific error that turns a swollen leg into a compartment syndrome.
17.2 How often — and why nobody can tell you
Evidence — an incidence range spanning a factor of seven
The 2019 narrative review reports limb ischaemia in "10 to 70%" of peripheral VA ECMO patients, and attributes the spread to "populations that are different in baseline characteristics, ECMO indications, cannulation techniques, limb ischemia definition."
Its specific cohorts show the same scatter: 8.6% (surgical cutdown with prophylactic distal perfusion), 12–33% despite a distal perfusion cannula, 14.7% in cardiogenic shock, 10.6% in ECPR. Downstream: fasciotomy in 8.3% of one cardiogenic shock cohort, and in 7 of 34 (21%) of the limb-ischaemia patients in another; amputation in 0.9% of one series and in 3 of 34 (8.8%) of limb-ischaemia patients.
The largest systematic estimate, a 2020 meta-analysis of 47 studies and 6,583 patients, pooled vascular complications overall at 29.5% (95% CI 23.6–35.9%) and limb ischaemia specifically at 12.6% (10.0–15.5%), with significant bleeding at 15.4% and cannula-site bleeding at 12.6%. Vascular complications were more than twice as likely on VA than VV support: OR 2.35 (1.87–2.96), p<0.0001.
A pooled analysis of 1,866 VA ECMO patients cited by the Red Book gives the most-quoted downstream figures: lower limb ischaemia 17%, compartment syndrome 10%, amputation 5%.
Certainty: low for any single figure; moderate for the ordering — ischaemia is common, compartment syndrome is a substantial minority of it, amputation is uncommon but not rare.
Why the numbers disagree is itself the clinically important point. There is no agreed definition of limb ischaemia on ECMO. A unit that records only limbs needing an operation will report 2%; a unit with a NIRS protocol that records every regional desaturation will report 30%. Both may be describing the same care. When comparing your unit's rate to a published one, compare the definitions first.
17.3 This is not really a limb problem
The reason to take an early dusky foot seriously is not primarily the foot.
Study | Survival with limb ischaemia | Survival without |
Single-centre analysis cited by the Red Book (2016), patients with any vascular complication | 18% discharged | 49% discharged |
Single-institution review across all ECMO modes, 229 patients (2022) | 32.5% in-hospital survival | 54% (log-rank p=0.023) |
Arterial-protocol cohort, 91 patients (2017) | 25% survival | 42% overall ECMO survival |
Peripheral VA cohort, 101 patients (2020) | 83% mortality | 57% mortality (p < 0.05) |
Ten-year single-centre series, 377 patients, 4-year follow-up (2024) | 46.8% four-year survival among those successfully decannulated | 65.1% (p=0.044); adjusted HR 1.80 (1.04–3.12), p=0.035 |
Danger — the association is consistent, large, and almost certainly not all causal
Five independent cohorts agree that patients who develop limb ischaemia die more often, and the ten-year series shows the effect persisting to four years after successful decannulation, with an 80% relative increase in mortality hazard.
It is tempting to read this as "prevent the ischaemia and you save the patient." Resist that reading. Limb ischaemia is caused by the things that also kill these patients — profound shock, high vasopressor doses, peripheral vascular disease, diabetes, low cardiac output. It is at least partly a marker of severity rather than a mechanism of death. This is exactly the confounding Chapter 15 §15.8 identified in the rescue-unloading data, and it recurs here.
But the practical conclusion survives either interpretation, and it is not the obvious one. If limb ischaemia is a marker, it is a marker you can read at the bedside, hours before the lactate moves. A leg that is failing on adequate circuit flow is telling you something about the whole circulation — usually that the vasoconstriction is too high, the output too low, or the patient too vasoplegic to perfuse anything peripheral.
Treat a newly ischaemic leg as a systemic finding as well as a local one. Ask what the vasopressor dose is doing before you ask what the cannula is doing.
17.4 Who is at risk — including two answers that surprise people
Risk factor | Evidence | What you can do about it |
Small common femoral artery | The 2019 review: the artery is smaller in women and younger patients, and "younger patients, who lack collateral circulation, seem to have smaller femoral arteries" — so youth is a double hit. A 139-patient series found a mean cannulated-side artery diameter of 0.82 versus 0.63 (units as printed; see the open items) separating no-ischaemia from ischaemia, p < 0.001 | Measure the artery before you cannulate (Chapter 12) and size to the vessel, not the target flow |
Cannula size and cannula-to-vessel ratio | Cannulae above 20 Fr are repeatedly associated with ischaemia. The 2019 review reports a lower incidence "when the relationship between body surface area and cannula size is greater than 11", and a protocol using 15 Fr cannulae achieving comparable support with fewer complications. A meta-analysis found a small-bore (under 17 Fr) arterial return cannula reduced limb ischaemia, OR 0.40 (0.24–0.65), p < 0.001 | "The smallest possible cannula should be preferred." But see the contradiction below |
Peripheral arterial disease | An independent risk factor in every source. In the ten-year series, 7.6% versus 3.0%, p=0.047; in a 139-patient series, 24% versus 9%, p < 0.001 | Screen at referral. Chapter 11 treats severe peripheral vascular disease as a relative contraindication to femoral VA support |
Diabetes | Independent in the 2019 review; in a 179-patient series, OR 4.338 (95% CI 1.193–15.772), p=0.026 — the largest single odds ratio in that model | Nothing acutely. It should raise the threshold for prophylactic distal perfusion and for how hard you look |
Vasopressor dose — the dominant modifiable factor | In the ten-year, 377-patient series the only independent predictor of acute limb ischaemia was vasopressor use, adjusted OR 6.8 (1.5–30.4), p=0.012. In a 179-patient series the peak vasoactive-inotropic score in the first 24 hours was independent, OR 1.054 (1.024–1.085), p < 0.001. A 139-patient series found ischaemia above a VIS of about 15.8 | This is the one you can change tonight. Chapter 13 §13.6.2: on VA ECMO, flow is the vasopressor. Coming down on noradrenaline is a limb intervention |
An intra-aortic balloon pump in the other groin | Independent in the 179-patient series, OR 1.526 (1.038–22.026), p=0.049 — note the very wide interval | Chapter 15's unloading decision has a limb cost. §17.11 |
Not being on aspirin | In the ten-year series, aspirin was protective: adjusted OR 0.52 (0.30–0.90), p=0.018; patients who developed ischaemia were less likely to have been taking it (40.2% versus 54.4%, p=0.013) | Hypothesis-generating only. A single retrospective analysis, and antiplatelet therapy on ECMO is Chapter 41's territory. It is recorded because it points at the thrombotic mechanism rather than the mechanical one |
Cannulation technique and setting | A 105-patient study found cannulation in the operating theatre reduced ischaemia (OR 0.25, 0.08–0.77, p=0.02) but increased bleeding (OR 2.65, 1.09–6.45, p=0.03); cutdown increased bleeding (OR 4.96, 2.32–10.61, p < 0.0001); and ultrasound guidance reduced both bleeding (OR 0.81) and in-hospital mortality (OR 0.41, 0.20–0.87, p=0.02) | Ultrasound guidance is the only item here with no trade-off. Chapter 12 |
Multiple cannulation attempts | Listed by the Red Book among the principal risk factors | The strongest argument for the most experienced available operator, first time |
Pitfall — the cannula-size rule is not as secure as it sounds, and one large series contradicts it
"Use the smallest cannula that will deliver the flow" is repeated everywhere in this literature, and the small-cannula meta-analysis supports it. But two findings sit awkwardly against it.
First, the ten-year, 377-patient series found arterial cannula size was not associated with acute limb ischaemia at all — nor was the concurrent use of a ventricular offloading device.
Second, and stranger, a 91-patient arterial-protocol study identified as a risk factor "ECMO cannula size of less than 20 Fr" — the opposite direction. That is most plausibly explained by selection rather than physiology (smaller cannulae go into smaller or more diseased arteries, and into patients cannulated in more difficult circumstances), but it is what the paper reports.
Certainty: low that cannula size is an independent causal driver; moderate that cannula-to-vessel ratio matters. Those are different claims, and conflating them is what produces the apparent contradiction. Size to the vessel you measured, and do not treat a small cannula as protection in a small artery.
17.5 Surveillance — both obvious monitors fail, for the same reason
This chapter's monitoring problem has the same root as Chapter 16's. VA ECMO reduces pulsatility, and the two bedside instruments for limb perfusion both depend on a pulse.
Physiology — why Doppler misleads on a leg the way pulse oximetry misleads on an arm
The Red Book is explicit: "Use of Doppler ultrasound may be deceptive as continuous ECMO flow may make distal pulses less pulsatile and therefore less detectable. NIRS monitoring can be advantageous in this setting as it monitors the difference between oxygenated and deoxygenated hemoglobin and does not require pulsatile blood flow."
The 2019 review quantifies the confound: distal peak systolic velocity is positively correlated with pulse pressure and negatively correlated with ECMO pump flow. So the Doppler signal degrades as a direct function of how much support you are giving — it is weakest exactly when the arterial cannula is largest relative to native flow, which is when the limb is most at risk.
Chapter 16 §16.4 made the same argument for pulse oximetry in the arm. The generalisation is worth stating once: on VA ECMO, every monitor that needs a pulse becomes least reliable as support increases. Regional oximetry does not need one.
The clinical consequence is measurable. A prospective study reported by the 2024 review found that only 50% of patients with loss of the Doppler pulse had clinical signs of limb ischaemia, and concluded that Doppler alone is "insufficient for VA ECMO distal limb perfusion assessment."
Pitfall — three of the six Ps are unavailable in a sedated patient
Both reviews teach the classical six Ps — pallor, pulselessness, paraesthesia, paralysis, pain and poikilothermia — and both immediately note the problem. The 2024 review states it plainly: assessment is "challenging when patients are sedated" because three of the six "are not assessable."
Pain, paraesthesia and paralysis are the early ones. What remains available in a sedated patient — pallor, pulselessness and a cold foot — are the late ones. The Red Book makes the same point: "other commonly described manifestations of limb ischemia such as pain and neurologic deficits may be difficult to identify due to patient sedation."
A structured examination several times per shift is still required — the 2019 review recommends exactly that — but do not build a surveillance strategy on findings your patient cannot report.
17.5.1 Near-infrared spectroscopy — the thresholds
Evidence — the numbers that make NIRS actionable
The physiological demonstration. In a 2008 study reported by the 2024 review, clamping the femoral artery dropped mean regional saturation from 61% to 38% (p=0.001), and placement of a distal perfusion cannula raised it to 71%. The signal tracks the thing you care about.
The intervention thresholds in widest use come from a protocol reported in both reviews: rSO₂ below 40%, or a fall of more than 25% from baseline, prompts placement or replacement of a distal perfusion cannula. In that series, 35% of VA ECMO patients met the criteria, and there was 100% restoration of NIRS values after placement or replacement, with no limb-related complications.
The ELSO figures, as reported by the 2024 review, are less aggressive and framed as targets rather than triggers: NIRS should be above 50%, preferably 60%, and the difference between the two legs should be less than 20%.
The predictive data. The 2019 review reports that every patient with clinical evidence of limb ischaemia had rSO₂ below 50% for longer than 4 minutes, with a calculated positive predictive value of 86%.
The timing data. A prospective study found NIRS detected ischaemia earlier than clinical assessment plus intermittent Doppler — mean time to perfusion 19.6 ± 21.4 versus 42.0 ± 69.0 hours — and no patient in the NIRS-guided group developed compartment syndrome, versus 13.9% of controls.
Certainty: moderate that NIRS detects limb malperfusion earlier and more reliably than clinical examination or Doppler; low for any specific threshold. The numbers come from small single-centre series with differing devices, and the very wide standard deviations in the timing study should be noticed.
Two practical requirements follow. First, baseline values on both legs before cannulation — Taha is explicit that this is what makes a later reading interpretable. Second, probes on both legs, not one.
Clinical pearl — the second probe is what turns a number into a diagnosis
A single NIRS value on the cannulated leg tells you that leg is desaturated. It does not tell you why, and the two commonest causes need opposite responses.
The Red Book records the manoeuvre that separates them: comparison of NIRS in both the cannulated and non-cannulated extremities differentiates "cannula-related obstruction" from "other sources of hypoperfusion (eg, high dose vasopressors)."
— Cannulated leg low, other leg normal → a local, mechanical problem. Distal perfusion, cannula position, thrombus.
— Both legs low → a systemic problem. Vasoconstriction, low flow, anaemia, hypothermia. A distal perfusion cannula will not fix this and adds a second arterial puncture to a patient who does not need one.
This is the limb equivalent of Chapter 16's paired post-oxygenator and right radial sample: one extra measurement, taken at the same moment, splits the differential in half.
17.5.2 Staging a threatened limb
Danger — the Doppler finding that means the limb is already lost
The 2019 review gives the staging that should govern urgency, and it hinges on which Doppler signal is missing:
— Loss of the arterial Doppler signal, venous signal present → a threatened limb. Salvageable. Act now.
— Absence of both arterial and venous Doppler signals → "irreversibly damaged and non-salvageable."
And the governing rule: "the longer the symptoms are present, the less likely the possibility of limb salvage."
The practical consequence for a sedated ECMO patient is severe. The findings that would ordinarily start the clock — pain, then paraesthesia, then paralysis — are the three that sedation removes (§17.5). Nobody will tell you when the clock started. That is the argument for continuous regional oximetry rather than intermittent examination, and it is the argument for acting on a NIRS trend rather than waiting for it to be confirmed clinically.
17.6 Perfusing the leg — technique, and the question of who gets one
17.6.1 The routes
Technique | How it is done | Advantages and limits |
Antegrade distal perfusion cannula into the superficial femoral artery — the default worldwide | A short 6–8 Fr armoured cannula into the proximal SFA, connected by short tubing with a male-to-male connector to the arterial limb, with a target flow of about 100 mL/min. The 2019 review reports sizes across the literature of 5–14 Fr, most commonly 6–8 Fr — central venous catheters and vascular introducer sheaths. Placed percutaneously under ultrasound or fluoroscopy, or by arteriotomy under direct vision. The wire is placed at the time of the main femoral cannulation (Chapter 12) | Direct control of distal flow; flexible sizing; percutaneous or surgical. Costs: an additional intervention, thrombosis risk in a low-flow segment, and extra procedure time |
Retrograde perfusion via a distal artery | Originally described through a 5 cm incision posterior to the medial malleolus with a 6–8 Fr cannula into the posterior tibial artery after arteriotomy; the dorsalis pedis and anterior tibial arteries are alternatives. Achieved flow 156 ± 82 mL/min | The answer when the antegrade route has failed or is inaccessible. Centre experience is usually limited. In the reported series, 8.3% (n=3) developed limb ischaemia when the cannula was inserted more than 6 hours after ECMO initiation — a timing signal that reinforces Chapter 14's deadline |
End-to-side graft (the "chimney") | A prosthetic graft sewn end-to-side to the common femoral artery and cannulated — originally a 10 mm PTFE graft; the 2019 review reports 6–8 mm Dacron or Hemashield | Bidirectional flow through one cannula, and a large cannula can be used without occluding the vessel — Taha notes it is "better suited to long-term venoarterial support." Costs: limited control over the ratio of retrograde to antegrade flow, unsuitable for emergencies, and wound oozing from high anastomotic pressures |
Femoro-femoral bypass | External: a 6 mm × 40 cm PTFE conduit run outside the body. Endovascular: a 5 Fr catheter with side holes and an end hole passed across the aortic bifurcation | Reserved for when antegrade femoral and retrograde tibial access have both failed. The external form is vulnerable to infection |
Bidirectional cannula | A single arterial cannula with an additional side hole at a 120-degree angled elbow, directing part of the flow distally. A 19 Fr device gave adequate distal flow in 14 of 15 patients checked with NIRS, with no ischaemic complications | One puncture, no inter-cannula thrombus risk. But less control over distal flow, which varies with circuit flow, anatomy and haemodynamics; limited sizes; and no reported experience during patient mobilisation (Chapter 59) |
Bilateral groin cannulation — artery in one groin, vein in the other | Not a perfusion technique but a configuration choice | The 2019 review considers it "preferable due to the reduction of vessel compression." A propensity-matched comparison showed less compartment syndrome and lower mortality with bilateral cannulation, though no overall reduction in limb ischaemia (§17.7) |
Danger — the distal perfusion cannula has three failure modes, and all are silent
1. It gets harder to place the longer you wait. The ECPR text is explicit: "placement becomes increasingly difficult over time as residual flow through the SFA decreases," and therefore "it is often preferred to place the distal perfusion cannula immediately after ECMO flow is initiated and the patient has stabilized." This is the mechanism behind Chapter 14's 4-hour rule and Chapter 12's wire-first rule — the vessel you need is being closed by the cannula you already placed.
2. It thromboses if the flow through it is too slow. "Slow flow through the sheath may produce thrombosis, so the sheath size should be maximized for the SFA" — typically 8–9 Fr are the largest used for this purpose. A too-small cannula is not a safer cannula here.
3. It kinks. "Legs bend and flex during patient transfers even if the patient is heavily sedated. Therefore, there is substantial risk of kinking the sheath leading to turbulent flow which also can induce thrombosis."
And it can simply be in the wrong place. A 229-patient institutional review found prophylactic distal perfusion cannulae were incorrectly positioned in 10.2% (7 of 68) — enough to obscure the analysis of whether they worked.
A distal perfusion cannula is not a device you place and forget. It is a device that requires the same daily verification as every other part of the circuit — flow through it, position, and the NIRS value it is supposed to be protecting.
17.6.2 Controversy 1 — prophylactic or selective?
Controversy 1 — Should every femoral VA ECMO patient get a distal perfusion cannula at cannulation?
The question. A prophylactic distal perfusion cannula means a second arterial puncture, in an anticoagulated patient, in a limb that in most cases would have been fine. A selective strategy means waiting for evidence of hypoperfusion — and risking that the vessel has closed by the time you need it (§17.6.1).
Position A — prophylactic, in everyone. This is the majority position and it has consistent meta-analytic support.
— 22 retrospective studies, 779 patients, 132 events: limb ischaemia 9.74% with a distal perfusion cannula versus 25.42% without; RR 0.41 (95% CI 0.26–0.65), p < 0.01, I²=28%. No mortality difference.
— 17 studies, 1,850 patients: RR 0.48 (0.37–0.62), p < 0.0001. No mortality difference.
— 17 studies, 2,040 patients (904 with a cannula): RR 0.49 (0.31–0.77), p=0.002, and directly on the timing question, prophylactic versus reactive placement RR 0.41 (0.24–0.71), p=0.02. No difference in mortality or bleeding.
— 22 studies: prophylactic placement OR 0.31 (0.21–0.47), p < 0.001.
— 47 studies, 6,583 patients: distal perfusion associated with lower odds of limb ischaemia and protective against mortality on meta-regression.
Individual cohorts point the same way: prophylactic placement in 55 of 91 patients with no subsequent ischaemia, against 33% ischaemia in those without; 2% versus 32% (p < 0.05) in another; and 2.1% versus 8.2% (p=0.047) in a 230-patient two-centre study, where fluoroscopy-guided simultaneous insertion was independently protective, OR 0.11 (0.01–0.98), p=0.048. The 2024 review states the resulting consensus flatly: current guidance recommends "distal limb perfusion to be initiated at the time of VA ECMO initiation in all patients."
Position B — selective, with a strict monitoring protocol.
— A 188-patient single centre with daily vascular surgical review and continuous regional saturation monitoring placed cannulae only on evidence of hypoperfusion. Across the whole cohort: major bleeding 3.7%, fasciotomy 3.7%, amputation 0%. The authors conclude a selective strategy "is safe and may obviate the risks of an additional arterial catheter."
— A 105-patient cohort found a distal perfusion cannula not associated with limb ischaemia (p=0.47), while ultrasound-guided cannulation reduced bleeding and in-hospital mortality.
— The 377-patient, ten-year series found the cannula not protective against acute limb ischaemia — but associated with fewer patients needing a vascular intervention, 20.1% versus 32.0%, p=0.009.
What the evidence actually shows. Every study on both sides is observational, and the two positions are not testing quite the same thing.
First, the meta-analyses pool studies in which "no distal perfusion cannula" usually also meant "no protocol." The selective-strategy centres are not doing nothing; they are doing continuous bilateral regional oximetry with daily vascular review, which §17.5 shows detects malperfusion around a day earlier than clinical assessment. The honest comparison is not cannula versus no cannula. It is cannula-in-everyone versus surveillance-plus-cannula-when-needed — and only one meta-analysis addressed timing directly.
Second, confounding by indication runs in opposite directions in the two literatures. In the selective centre, 30-day mortality was highest in the upfront-cannula group (56% versus 19% delayed and 22% none, p < 0.001) — which almost certainly reflects who was chosen for upfront placement, not harm from it. Read as evidence against prophylaxis it would be badly misleading.
Third, the two positions agree on more than they disagree. No study on either side shows a mortality benefit from the cannula itself. What differs is whether you can reliably detect the 10–30% who need one in time — and that depends entirely on whether your unit actually runs the surveillance.
Certainty: moderate that a prophylactic distal perfusion cannula reduces limb ischaemia. Low that it changes mortality. Low that a selective strategy is equivalent, and it is conditional on a protocol most units do not have.
Where practice actually sits. Most units place prophylactically, and some place bilaterally when a second arterial device is involved (§17.11). Selective strategies exist in centres with daily vascular surgical involvement.
What would resolve it. A randomised comparison of prophylactic placement against protocolised NIRS-triggered placement, powered for limb ischaemia. Every meta-analysis in this area ends by asking for one; none is registered.
What to take to the bedside. Place prophylactically unless your unit can honestly answer yes to three questions: are both legs monitored continuously, is there a written trigger threshold, and can someone place a cannula within the hour at three in the morning? If any answer is no, the selective strategy is not the strategy those papers describe. And whichever you choose, the wire goes in at the time of the main cannulation — that decision is separate from, and earlier than, the decision to perfuse.
17.7 The forgotten half — venous congestion
Almost all the attention, all the devices and all the meta-analyses concern arterial inflow. The 2024 review is blunt about what that misses.
Physiology — a limb can be adequately perfused and still die
"Venous obstruction by the venous cannula, leading to venous stasis, and hence reduced arterial perfusion, may be equally dangerous."
The mechanism is straightforward and it is the same one that produces compartment syndrome anywhere: a large drainage cannula occupies the femoral vein; outflow falls; interstitial and venous pressure rise; the arteriovenous pressure gradient across the limb narrows; capillary perfusion falls even though the artery is patent. Taha adds the specific configuration that causes it — cannulating the femoral artery and vein on the same side — and notes that arterial or venous outflow obstruction combined with hyperperfusion makes compartment syndrome "a serious condition."
Taha also records the quieter consequence: lower limb deep vein thrombosis, if the drainage cannula occupies the entire vein.
The trap this sets is specific and dangerous. Adding arterial inflow to a limb that cannot drain raises compartment pressure faster. The 2024 review warns directly that excess distal flow "increases the risk for venous congestion leading to severe tissue damage to the leg."
Venous strategy | What it does | Evidence |
Bilateral cannulation — artery and vein in opposite groins | Removes the compression entirely by not putting two large tubes in one femoral sheath | A large retrospective propensity-matched comparison showed less compartment syndrome and lower mortality, but no overall reduction in limb ischaemia. The 2024 review recommends bilateral cannulation in addition to selective arterial perfusion |
Facilitated venous drainage — a second draining cannula in the distal femoral vein | Actively decompresses the limb. Described as a four-cannula configuration: arterial, venous, distal perfusion, and a 16 Fr secondary extracting venous cannula draining the distal femoral vein, achieving 4–6 L/min patient flow with 1–2 L/min of distal arterial flow | Small series. An 11 Fr introducer in the distal femoral vein has also been reported to treat limb oedema from venous stasis |
Duplex assessment of the vein | Identifies stasis rather than assuming it | The suggested signs are continuous flow with a small difference between maximum and minimum velocity in the femoral vein on the cannulated side, and a large vein-diameter ratio. Certainty: very low — the review describes these as pilot findings in a small cohort |
Clinical pearl — the swollen leg and the pale leg are different diseases
Both present as a cool, dusky limb on the cannulated side, and the reflex — check the distal perfusion cannula — is right for only one of them.
— Pale, empty, collapsed veins, low NIRS on the cannulated side only → arterial. Distal perfusion, cannula position, thrombus.
— Swollen, tense, mottled, engorged, with a measurable diameter difference → venous. More arterial inflow makes it worse. Consider drainage: repositioning, a second drainage cannula, or moving the vein to the other groin.
If the leg is bigger than the other one, think about outflow before inflow.
17.8 Treatment — the conservative ladder comes first, and it is mostly not surgical
The 2019 review's conservative ladder deserves to be better known, because five of its six steps cost nothing and can be started before anyone is called.
Conservative step | Why |
Reduce or discontinue vasopressors | The dominant modifiable risk factor (§17.4). On VA ECMO you can raise flow instead (Chapter 13 §13.6.2) — but check the aortic valve and the right radial gas before you do (Chapters 15, 16) |
Optimise volume status and haemoglobin oxygen transport | A limb at the end of a partially obstructed artery is living on marginal delivery. Anaemia and hypovolaemia are on the same side of that equation |
Optimise peripheral temperature | A cold limb is a vasoconstricted limb. The circuit heat exchanger is under your control |
Administer a peripheral vasodilator through the distal perfusion cannula | Regional delivery, systemic dose avoided. Rarely taught; explicitly listed in the review |
Maintain anticoagulation at the highest therapeutic level | Two of the six mechanisms are thrombotic. This is the step that competes hardest with everything else the patient needs (Chapter 45) |
Place or replace the distal perfusion cannula | If one is absent, place it. If one is present and the NIRS has fallen, assume it has thrombosed, kinked or migrated until proven otherwise (§17.6.1) |
Invasive step | When |
Remove and reposition the arterial cannula — contralateral limb, subclavian, or aorta | When the cannula itself is the obstruction and the limb is threatened. Note that this is the same manoeuvre Chapter 16 §16.10 recommends for differential hypoxaemia — one decision can solve both problems |
Fogarty catheter embolectomy | Thromboembolic occlusion. Must be distinguished from dissection-induced occlusion (§17.10) |
Arterial repair — suture, or patch angioplasty with bovine pericardium | Cannulation injury, or at decannulation (Chapter 18) |
Fasciotomy | Established compartment syndrome — a tense calf with pain and paraesthesia. The ISCCM manual adds the step everyone forgets: after fasciotomy, anticoagulation usually has to be adjusted to prevent bleeding from the wounds |
Amputation | Irreversible damage. In pooled data this is around 5% of VA ECMO patients overall, and 8.8% of those with limb ischaemia in one series |
17.9 The other end of the patient — upper-limb hyperperfusion
Chapter 16 gives a good reason to move the arterial return to the subclavian or axillary artery. Chapter 12 recorded the price. This is where it is paid.
Danger — the same operation that saves the brain can cost the arm
Taha describes the mechanism: "Hyperperfusion states can result in swollen and hyperemic limb in arterial cannulation in some cases where the axillary or subclavian artery was cannulated, or if the femoral artery and vein are cannulated on the ipsilateral side. The presence of arterial or venous outflow obstruction, in addition to hyperperfusion, can make compartmental syndrome a serious condition in which high compartment pressure may lead to limb ischemia."
The ISCCM manual is consistent: "The problem with axillary or subclavian artery cannulation may be distal limb hyperemia. It can lead to compartment syndrome and may require fasciotomy. This technique is more invasive and there is also a higher risk of injury to the vessels and nerves of the arm."
Three points follow.
1. The pathway to injury is the same as in the congested leg — high compartment pressure — reached from the opposite direction. A hyperperfused arm and an obstructed leg both end in fasciotomy.
2. The arm has less room. Brachial plexus injury is a risk unique to this site, and the review literature notes that the end-to-side graft is the standard mitigation, allowing flow to divide between the arm and the aorta rather than being forced entirely down the arm.
3. Monitor it the same way. Everything in §17.5 applies: baseline bilateral values before cannulation, and continuous regional oximetry on both arms. The arm gets less attention than the leg for no defensible reason.
Certainty: low — this is textbook description and case-level reporting; no comparative series of upper-limb complication rates was retrieved.
17.10 Vascular injury that is not ischaemia
Lesion | How it presents | Management |
Arterial dissection | Most are asymptomatic. When they are not, they cause occlusion — and Taha's key point is that dissection-induced occlusion must be distinguished from thromboembolic occlusion, because the treatments differ | Symptomatic dissection: relocate the cannula and stent |
Access-site pseudoaneurysm | Painful pulsatile groin swelling, confirmed on ultrasound. Untreated it can cause arterial occlusion, rupture, haematoma or infection | Ultrasound defines the neck. Narrow-necked lesions can be treated with thrombin injection; wide-necked lesions need surgery. Local compression is the immediate measure; endovascular repair is an alternative for small lesions |
Retroperitoneal haemorrhage and large internal haematoma | The one to fear, because there is nothing to see. An unexplained fall in haematocrit with haemodynamic compromise should prompt a search for vascular injury, confirmed on imaging. Anticoagulation means minor vascular injuries produce major haematomas | Correct anticoagulation, transfuse. Endovascular embolisation if conservative measures fail; open surgery rarely |
Cannulation-site bleeding | The commonest bleeding source of all. Pooled prevalence 12.6% | Taha's sequence: verify anticoagulation level, platelets, prothrombin time and fibrinogen; reduce anticoagulation targets; topical haemostatic agents. Prevention is technical — avoid large incisions and over-dilatation so that the cannula fits tightly. Surgical re-exploration if that fails (Chapters 36, 45) |
Cannula-site infection | Cellulitis, swelling or discharge. Groin infections are particularly likely in obese and malnourished patients | Antibiotics promptly, surgery if needed. If bacteraemia persists, the circuit itself may need replacing — its surface area can prevent eradication (Chapter 58) |
Danger — the complication that appears after the patient has gone home
The ISCCM manual describes a category this book has not yet covered anywhere: late vascular complications — arterial stenosis at the former cannulation site.
Its independent predictors are named: technical problems during explantation, and a history of peripheral vascular disease. The treatments are all vascular-surgical: percutaneous transluminal angioplasty, femorofemoral crossover bypass, iliofemoral bypass, or thromboendarterectomy.
It also puts the overall figure on injury at cannulation and removal: dissection, pseudoaneurysm and retroperitoneal bleeding occur in 7–14% of patients across placement and removal combined.
Two consequences. First, decannulation is a vascular operation, not the end of one (Chapter 18) — and how it is done determines a complication that may not declare itself for months. Second, an ECMO survivor with new claudication has a diagnosis until proven otherwise, and the unit that cannulated them is the one most likely to think of it. Certainty: low — a textbook account without retrieved primary data on incidence or timing.
17.11 Two devices, one limb
Chapter 15 may put a micro-axial pump in the other groin. Chapter 16 may add a return limb. The arithmetic of the limb changes when it does.
Evidence — what the second arterial device costs, and one protocol's answer
The propensity-matched international cohort in Chapter 15 §15.7.1 reported, in patients unloaded with a micro-axial pump alongside VA ECMO, access-site ischaemia 21.6% versus 12.3% and abdominal compartment syndrome 9.4% versus 3.7%.
One centre's response, reported in 2024: bilateral 6 Fr antegrade superficial femoral distal perfusion cannulae in every combined case, regardless of vessel diameter or of flow past the device sheath at the time of placement. In 49 consecutive runs of more than 24 hours, bilateral cannulae were placed in 42 patients, with no limb ischaemia in that group. Of the 7 who had no distal perfusion cannula on the device side, 3 developed limb ischaemia — one needing device removal, two needing fasciotomy; none required amputation.
Certainty: very low — a single-centre retrospective series of 49, reported as a conference abstract. It is included because it addresses a question the meta-analyses do not ask at all.
What is reasonably solid underneath it: a second large arterial sheath in the second groin means both legs now carry the risk that §17.4 describes for one, and the surveillance in §17.5 must be bilateral by default. If a decision in Chapter 15 or 16 puts a second arterial device in, the distal perfusion question must be settled for that limb at the moment of insertion — not at the next ward round.
17.12 The errors that recur
Error | Correction |
Using an absent Doppler pulse as the test of limb perfusion | Doppler velocity falls as circuit flow rises. It is weakest when the limb is most at risk, and only half of patients who lose the pulse have clinical ischaemia |
Monitoring one leg | A single value cannot separate a cannula problem from a vasopressor problem. The second probe is what makes the first interpretable |
Reading NIRS without a pre-cannulation baseline | The most widely used trigger is a fall of more than 25% from baseline. Without a baseline you have thrown that trigger away |
Relying on pain and paraesthesia in a sedated patient | Three of the six Ps are unavailable, and they are the early three |
Adding arterial inflow to a swollen, congested leg | That is venous obstruction, and more inflow raises compartment pressure faster |
Treating an ischaemic leg as purely local | Vasopressor dose was the only independent predictor in the largest long-term series. Check the noradrenaline before the cannula |
Assuming an existing distal perfusion cannula is working | They thrombose, kink with leg movement, migrate, and in one series were misplaced in 10.2%. A falling NIRS with a cannula in place means verify the cannula |
Deferring distal perfusion until it is needed | The superficial femoral artery becomes progressively harder to access as flow past the main cannula falls. The wire goes in at the main cannulation |
Choosing the smallest possible distal perfusion cannula | Too little flow through the sheath thromboses it. Maximise it for the vessel — typically 8–9 Fr |
Treating a small arterial cannula as protection in a small artery | Cannula-to-vessel ratio is the variable, not absolute size; one large series found no association with size at all |
Forgetting the arm after moving the return to the subclavian artery | Hyperperfusion ends in the same fasciotomy from the opposite direction. Baseline and monitor both arms |
Attributing an unexplained haematocrit fall to haemolysis or bleeding elsewhere | Retroperitoneal haemorrhage from a minor vascular injury is invisible and common under anticoagulation. Image it |
Treating decannulation as the end of the vascular story | Late arterial stenosis at the cannulation site is predicted by technical difficulty at explantation. New claudication in a survivor is a diagnosis |
Placing a second arterial device without settling the distal perfusion question for that limb | Both legs now carry the risk. Decide at insertion |
17.13 Key points
- Six mechanisms threaten the limb — arterial obstruction, cannulation injury, loss of pulsatile flow, thromboembolism, venous stasis, and systemic vasoconstriction. A distal perfusion cannula addresses two of them.
- Reported incidence spans 10–70% because there is no agreed definition. Pooled estimates put limb ischaemia around 12.6% and all vascular complications around 29.5%, with vascular complications more than twice as likely on VA than VV support.
- Downstream: compartment syndrome in roughly 10% and amputation in roughly 5% of VA ECMO patients in pooled data.
- Limb ischaemia predicts death, consistently across five cohorts and persisting to four years after decannulation — but it is at least partly a severity marker. Read it as a systemic finding you can see early.
- Vasopressor dose is the dominant modifiable risk factor — the only independent predictor in the largest long-term series. On VA ECMO you have an alternative lever: flow.
- Small arteries, young patients, women, peripheral arterial disease and diabetes raise risk. Cannula-to-vessel ratio matters more than absolute cannula size, and one large series found no association with size at all.
- Doppler degrades as circuit flow rises, because distal peak systolic velocity tracks pulse pressure. It is the same failure that breaks pulse oximetry in Chapter 16 — on VA ECMO, every monitor that needs a pulse fails as support increases.
- Three of the six Ps are unavailable in a sedated patient, and they are the early three.
- Regional oximetry does not need a pulse. Widely used triggers are rSO₂ below 40% or a fall of more than 25% from baseline; ELSO targets are above 50%, preferably 60%, with less than 20% difference between legs. Clinical ischaemia was preceded by rSO₂ below 50% for more than 4 minutes with a positive predictive value of 86%.
- Take baseline values on both legs before cannulation, and monitor both afterwards. The second probe separates a cannula problem from a vasopressor problem.
- Loss of the arterial Doppler signal with the venous signal preserved is a threatened limb. Loss of both is a limb that cannot be saved. Sedation hides when the clock started.
- The distal perfusion cannula gets harder to place every hour, thromboses if flow through it is too slow, kinks when the leg moves, and is sometimes simply in the wrong place. It requires daily verification.
- Every meta-analysis supports prophylactic placement — risk ratios of 0.41 to 0.49, and prophylactic beats reactive at RR 0.41 (0.24–0.71). No study on either side shows a mortality benefit.
- A selective strategy is defensible only with continuous bilateral monitoring, a written trigger and an operator available out of hours. Otherwise it is not the strategy those papers describe.
- Venous congestion is the forgotten half. A leg can be adequately perfused and still die from obstructed outflow, and adding arterial inflow makes it worse. Bilateral groin cannulation reduces compartment syndrome.
- The conservative ladder comes first and is mostly free: cut vasopressors, optimise volume and haemoglobin, warm the limb, give a vasodilator through the distal perfusion cannula, and anticoagulate at the top of the therapeutic range.
- After fasciotomy, anticoagulation usually has to be reduced — the wounds bleed.
- Subclavian or axillary return trades ischaemia for hyperperfusion, and reaches compartment syndrome from the opposite direction. Monitor the arm the way you monitor the leg.
- Unexplained haematocrit fall with instability is a vascular injury until imaged. Anticoagulation turns minor injuries into major haematomas.
- Decannulation is a vascular operation, and technical difficulty at explantation predicts late arterial stenosis. New claudication in an ECMO survivor is a diagnosis.
[VERIFICATION REQUIRED] — open items in this chapter
- Two reviews were retrieved in full and carry most of the content: the 2019 Critical Care narrative review (DOI 10.1186/s13054-019-2541-3) and the 2024 Perfusion review (DOI 10.1177/02676591241236650). A first-author discrepancy affects the first of these: a bibliographic index returns it under Bonicolini as first author, while the retrieved article's own citation line gave Lorusso. The DOI, journal, year, volume and content agree. No author list is asserted — the same rule applied to the JACC 2019 discrepancy in Chapter 15.
- All five meta-analyses and all cohort studies in §§17.3, 17.4 and 17.6.2 were read as structured abstracts, not full text. Sample sizes, effect estimates and confidence intervals are as reported in those abstracts; author lists, volumes and page numbers are not asserted.
- A femoral artery diameter is reported in one source as "0.82 versus 0.63" with a threshold of "≤6.3 cm". Those units cannot both be right, and 6.3 cm is not a plausible common femoral artery diameter — it is almost certainly millimetres. The figures are reproduced as printed with the discrepancy flagged, not silently corrected, and should not be used as a threshold without checking the primary source.
- A 2020 meta-analysis reports that distal perfusion was "associated with lower odds of limb ischemia" while quoting an odds ratio above 1. The direction of the estimate and the direction of the stated conclusion are inconsistent in the source. Only the qualitative conclusion is used here, and the numeric estimate is deliberately not quoted.
- A NIRS criterion for distinguishing cannula-related obstruction using the difference between the cannulated and non-cannulated leg appears in the 2019 review, but the direction of the threshold as extracted is internally inconsistent. The bilateral-comparison principle is used — it is independently supported by the Red Book and by ELSO's "less than 20% difference" target — but the specific delta figure from that review is not reproduced.
- The "body surface area to cannula size greater than 11" relationship is quoted as printed; the units and derivation were not established.
- The 8.8% limb ischaemia rate for retrograde tibial perfusion placed beyond 6 hours rests on 3 patients.
- The bilateral distal perfusion protocol in combined device cases (§17.11) is a single-centre conference abstract of 49 runs.
- The aspirin association (§17.4) is one retrospective analysis and is presented as hypothesis-generating only. Do not start antiplatelet therapy on this basis — Chapter 41 owns that decision.
- The intra-aortic balloon pump odds ratio of 1.526 (1.038–22.026) has an interval so wide as to be nearly uninformative; it is reported for completeness.
- The late vascular complication account, the 7–14% figure for injury at placement and removal, and the post-fasciotomy anticoagulation point are from the ISCCM manual without retrieved primary data.
- Upper-limb hyperperfusion (§17.9) has no retrieved comparative data at all — it is textbook description. The claim that it is under-monitored relative to the leg is this book's observation, not a published finding.
- The six-mechanism framing in §17.1, and the argument in §17.6.2 that the meta-analyses compare cannula versus no protocol rather than cannula versus surveillance, are this book's reasoning. The underlying facts are cited; the synthesis is not.
- Red Book, ISCCM, Taha and ECPR and Resuscitative ECMO are cited by chapter; editions, editors, years and page numbers are not confirmed.
Cross-references
- Chapter 11 — VA ECMO: Indications and Patient Selection: peripheral vascular disease as a relative contraindication
- Chapter 12 — VA ECMO Cannulation: measuring the vessel, sizing the cannula, ultrasound guidance, and the distal perfusion wire before the arterial cannula
- Chapter 13 — VA ECMO Haemodynamics: §13.6.2 flow as the vasopressor — the lever that makes §17.8's first step possible
- Chapter 14 — Initial VA ECMO Management: the 4-hour distal perfusion deadline, and why it exists
- Chapter 15 — LV Distension and LV Unloading: the second arterial device, and its access-site cost
- Chapter 16 — Differential Hypoxaemia: the same pulsatility failure in a different monitor, and the subclavian return that trades one limb problem for another
- Chapter 18 — VA ECMO Weaning and Decannulation: decannulation as a vascular operation, and the origin of late stenosis
- Chapter 34 — Cannula Problems: malposition, kinking and migration on the circuit side
- Chapters 36 and 45 — Bleeding and Bleeding Management: cannulation-site and retroperitoneal haemorrhage
- Chapter 41 — Antiplatelet therapy on ECMO: where the aspirin signal in §17.4 belongs
- Chapter 58 — Infection Prevention and Treatment: cannula-site infection and circuit seeding
- Chapter 59 — Mobilization and Rehabilitation: moving a patient whose distal perfusion cannula kinks when the leg bends
- Chapter 80 — ECMO + Impella: the combined-device limb problem in full
References
Retrieved in full
- Limb ischemia in peripheral veno-arterial extracorporeal membrane oxygenation: a narrative review of incidence, prevention, monitoring, and treatment. Critical Care. 2019;23:266. DOI 10.1186/s13054-019-2541-3. [VERIFICATION REQUIRED] — first author given as Bonicolini by one index and Lorusso by the retrieved citation line; no author list asserted.
- Simons J, et al. Evolution of distal limb perfusion management in adult peripheral venoarterial extracorporeal membrane oxygenation with femoral artery cannulation. Perfusion. 2024;39(1_suppl):23S–38S. DOI 10.1177/02676591241236650.
Syntheses — read as structured abstracts
- Juo Y, et al. Efficacy of Distal Perfusion Cannulae in Preventing Limb Ischemia During Extracorporeal Membrane Oxygenation: A Systematic Review and Meta-Analysis. Artificial Organs. 2017. DOI 10.1111/aor.12942.
- Gouchoe DA, et al. Does Size Matter? The Effect of Size of Distal Perfusion Catheter on Acute Limb Ischemia: A Meta-Analysis. ASAIO Journal. 2024. DOI 10.1097/MAT.0000000000002178.
- Marbach J, et al. Strategies to reduce limb ischemia in peripheral venoarterial extracorporeal membrane oxygenation: A systematic review and meta-analysis. International Journal of Cardiology. 2022. DOI 10.1016/j.ijcard.2022.04.084.
- Jia D, et al. Vascular Complications of Extracorporeal Membrane Oxygenation: A Systematic Review and Meta-Regression Analysis. Critical Care Medicine. 2020. DOI 10.1097/CCM.0000000000004688.
- Zhang X, et al. Distal perfusion cannulae in preventing limb ischemia in venous-arterial extracorporeal membrane oxygenation: a meta-analysis. Zhonghua Wei Zhong Bing Ji Jiu Yi Xue. 2020. DOI 10.3760/cma.j.cn121430-20200920-00639.
Cohorts — read as structured abstracts
- Lamb KM, et al. Arterial protocol including prophylactic distal perfusion catheter decreases limb ischemia complications in patients undergoing extracorporeal membrane oxygenation. Journal of Vascular Surgery. 2017. DOI 10.1016/j.jvs.2016.10.059.
- Nejim B, et al. Acute Limb Ischemia In Patients On Veno-Arterial Extracorporeal Membrane Oxygenation Support: A Ten-Year Single-Center Experience. Annals of Vascular Surgery. 2024. DOI 10.1016/j.avsg.2024.11.002. Source for the vasopressor and aspirin findings and the 4-year mortality hazard.
- Buda KG, et al. Routine Versus Selective Distal Perfusion Catheter Use in Venoarterial Extracorporeal Membrane Oxygenation. ASAIO Journal. 2024. DOI 10.1097/MAT.0000000000002264. The selective-strategy position.
- Son A, et al. Limb ischemia and bleeding in patients requiring veno-arterial extracorporeal membrane oxygenation. Journal of Vascular Surgery. 2020. DOI 10.1016/j.jvs.2020.05.071.
- Blakeslee-Carter J, et al. Vascular Complications Based on Mode of Extracorporeal Membrane Oxygenation. Journal of Vascular Surgery. 2022. DOI 10.1016/j.jvs.2022.01.078. Source for the 10.2% misplacement figure.
- Hu S, et al. Limb Ischemia Complications of Veno-Arterial Extracorporeal Membrane Oxygenation. Frontiers in Medicine. 2022. DOI 10.3389/fmed.2022.938634. Diabetes, IABP and vasoactive-inotropic score.
- Jang W, et al. Fluoroscopy-guided simultaneous distal perfusion as a preventive strategy of limb ischemia in patients undergoing extracorporeal membrane oxygenation. Annals of Intensive Care. 2018. DOI 10.1186/s13613-018-0445-z.
- Hanley S, et al. Distal perfusion cannulae reduce ECMO-related limb ischemia. International Angiology. 2020. DOI 10.23736/S0392-9590.20.04408-9.
- Yen C-C, et al. Identifying the Risk Factor and Prevention of Limb Ischemia in Extracorporeal Membrane Oxygenation with Femoral Artery Cannulation. The Heart Surgery Forum. 2018. DOI 10.1532/hsf.1824. [VERIFICATION REQUIRED] — arterial diameter units as printed are internally inconsistent.
- Hart JP, et al. Vascular Complications in Extracorporeal Membrane Oxygenation — A Narrative Review. Journal of Clinical Medicine. 2024. DOI 10.3390/jcm13175170. Source for the six-mechanism list.
- Whitmore S, et al. Standard Placement of Bilateral Distal Perfusion Catheters to Prevent Limb Ischemia During Femoral Veno-Arterial Extracorporeal Membrane Oxygenation Plus Impella CP. ASAIO Journal. 2024. DOI 10.1097/01.mat.0001069872.69909.37. Conference abstract, 49 runs.
Textbooks
- ELSO Red Book, 6th edition, Chapter 26 (cannulation complications) and Chapter 33 (vascular and haematologic complications of adult cardiac ECLS — the pooled 17%/10%/5% figures, the Doppler-versus-NIRS argument, the bilateral NIRS comparison, and the 18%-versus-49% discharge data). [VERIFICATION REQUIRED] — page numbers not confirmed.
- ISCCM Manual of RRT and ECMO in ICU, Chapter 41 (limb ischaemia, vascular injury at placement and removal, pseudoaneurysm neck anatomy and thrombin injection, post-fasciotomy anticoagulation, and late vascular complications). [VERIFICATION REQUIRED].
- Taha AR, Caridi-Scheible M, Leiendecker E, et al. ECMO: A Practical Guide to Management, Chapter 7 (vascular injury, leg ischaemia, bilateral NIRS baselines, end-to-side graft, lower limb DVT, hyperperfusion compartment syndrome, cannulation-site bleeding and infection). [VERIFICATION REQUIRED].
- Shinar Z, Badulak J, eds. ECPR and Resuscitative ECMO, Chapter 5 (distal perfusion cannula technique, the difficulty of patient selection under changing vascular tone, the argument for placing one in all patients, sheath sizing and the kinking risk). [VERIFICATION REQUIRED].
Chapter status
Drafted and audited 10 September 2026. Ten-pass quality control completed: clinical, physiology, evidence, citation, numerical, safety, contradiction, redundancy, bedside utility and literature-currency passes.
The numerical audit was the substantial one here and it found three separate problems in published sources: a femoral artery diameter reported in units that cannot be right, an odds ratio whose direction contradicts the conclusion drawn from it, and a bilateral NIRS threshold whose stated direction is internally inconsistent. All three are flagged and none is silently corrected; in two cases the numeric value is deliberately not reproduced.
The contradiction audit surfaced the prophylactic-versus-selective disagreement (§17.6.2) and a subsidiary one on cannula size (§17.4), where one large series found no association and another reported the opposite direction. The proposed reconciliation — that the meta-analyses compare a cannula against no protocol rather than against surveillance — is labelled as this book's reasoning.
The redundancy audit removed cannula sizing and site selection, which belong to Chapter 12, and the 4-hour deadline, which belongs to Chapter 14; both are referenced rather than restated.