The Pre-brief
After a well deserved vacation, you may come back on service to a patient with ARDS on a VV-ECMO circuit. If you need a refresher on what VV-ECMO is, or the decision making process involved with starting that type of salvage, please stop now and read our prior article [link]. The focus is no longer on the ICU to start a heroic effort; now it’s the day-to-day maintenance of ECLS that takes place to ultimately support and wean. While less dramatic than the “going on pump” moment, the daily maintenance, weaning, and safety measures are exactly what makes ECMO centers succeed in lung rescue.
Sometimes the circuit needs troubleshooting. One issue clinicians will encounter is progressive or sudden hypoxia in a previously stable VV-ECMO patient. While this complaint deserves a comprehensive approach to diagnosis and treatment, recirculation is a common cause of hypoxia, and it’s crucial to be able to recognize it.
What is Recirculation?
Unlike VA-ECMO, VV-ECMO drains and returns blood into the same venous system. That makes it possible for some amount of freshly oxygenated blood returning to the patient to bypass the right atrium and instead go directly into the drainage cannula (see figure 1). Recirculated blood reduces the efficiency of the circuit, as any amount of recirculation does not contribute to systemic oxygenation (1-4). While more commonly an issue for 2-cannula configurations, recirculation can also occur with single-cannula strategies if injection ports are inappropriately positioned. Although allowable at low levels, significant recirculation will cause dangerously low oxygen delivery despite a functioning ECMO circuit.

Recognizing Recirculation
Often the first noted indication of recirculation is a progressively decreasing pulse-ox/ peripheral O2 saturation, usually below an SpO2 of 86% (1). It’s important to remember that VV-ECMO patients never really need to be maintained at SpO2 levels higher than 92%, and 88% is usually acceptable(again, see our previous article if needed). The second, often confirmatory, sign of recirculation is an increasing SpreO2, i.e. pre-oxygenator oxygen saturation. This value is taken from a blood gas sampled from blood immediately before entering the membrane lung. Often the first impulse for VV-ECMO patients that have worsening hypoxia is to trial an increased pump flow (thereby increasing oxygen delivery). With significant recirculation, higher pump speeds/ flow often results in more recirculation rather than improving it (2, 3). While there may be a “sweet spot” where recirculation is at its lowest, increasing flow beyond that point rarely improves anything.
The percent of recirculation is commonly calculated with the following equation:
Recirculation (%) = (SpreO2 – SvO2) / (SpostO2 – SvO2) × 100
SpreO2 = oxygen saturation of blood entering oxygenator (“pre-membrane gas”)
SpostO2 = oxygen saturation of blood exiting the oxygenator (“post-membrane gas”)
SvO2 = oxygen saturation of venous blood returning to the vena cavae just before being drained by the ECMO circuit
Of all of these values, the most controversy revolves around SvO2; it is nearly always an estimate. It can be approximated by taking a VBG off the distal port of a CVC, though this can be inaccurate given the distal tip’s position relative to the re-infusion cannula. Another way to measure SvO2 is by performing a brief “weaning” trial. The sweep gas is turned off and the ventilator is used to achieve the same SpO2 level that ECMO support accomplished. Under these circumstances SpreO2 (either by blood gas or non-invasive measurement) effectively becomes SvO2. While more reliable, this is obviously more difficult to do in unstable patients (3).
Still other methods exist, including use of oxygen content fractions and dilutional ultrasound. None of these methods are often easy to apply at the bedside; however, and while used in vitro, are not routinely applied clinically (4).
Though no standardized threshold exists, it seems generally accepted that recirculation beyond 20-30% is nearly always clinically important, and a SpreO2 of <75% speaks to essentially negligible recirculation (1, 4).
What Causes Recirculation?
The differential for worsening recirculation breaks down to:
- Cannula migration: has the patient moved or the cannula shifted such that the return cannula is no longer aimed at the RA/ tricuspid valve?
- Excessive flow rates
- Abnormal pressure differentials causing increased resistance within the RA (i.e. elevated intra-cardiac, intrathoracic, or intra-abdominal pressures).
Systematic interrogation of each of these causes will help solve the problem quickly.
Cannula Position:
- Evaluate with CXR and/or fluoroscopy or TEE to make sure cannulae have not migrated and that flow is appropriately directed
- For 2-cannula setups, increase the distance between inflow and outflow cannulae (optimal distance in adults is often around 15 cm).
- Consider placement of a separate drainage cannula, allowing for a similar effective pump flow at lower pump speeds. The physics here dictate that if there are 2 drainage sites, then total drainage flow will be the same, but the amount of negative pressure needed at either site will be considerably reduced whereas the flow and pressure of oxygenated blood returning to the patient will remain the same.
- Consider transition to a dual-lumen or more centrally located catheter (e.g. Avalon or Protek Duo).
Excessive Pump Speed/ Circuit Flow:
VV-ECMO supplements native venous circulation with enough oxygen content to prevent organ failure and promote healing, and that oxygen content/ delivery is proportional to circuit flow. There is always some degree of circuit flow that misses the right atrium, mostly due to positioning or the interaction of return flow and native circulation. At normal or lower flow settings essentially all machine-oxygenated blood returns directly to the right atrium. As flow increases, however, a portion of circuit flow may run against native circulation in the caval system (e.g. Figure 1). Where exactly this threshold of diminishing returns exists is a patient-specific phenomena, and in appropriate configurations it’s rarely an issue. However, it is important to be aware that a ceiling for pump flows nearly always exists, especially in 2-cannula setups.
Worsening physiologic pressure differentials:
VV-ECMO patients are prone to multiple and spontaneous complications. If there is a physical obstruction to blood flow, such as elevated resistance in the pulmonary vasculature or higher pressures within the chest, this can also cause recirculation. In these cases, early diagnosis and management may be even more critical. If flow limitation becomes extreme enough, recirculation will proceed to complete loss of venous drainage and stop pump flow entirely.
- Make sure the patient is calm and not having large degrees of high peak pressures on the ventilator/sedated well. Pharmacologic paralysis may help take patient effort out of the equation here, so consider it early, at least temporarily as a diagnostic maneuver.
- Check for/address events that could be elevating intrathoracic or intra-atrial pressures or decreasing chest wall compliance:
- Tension pneumothorax
- Pericardial Tamponade
- Lobar collapse or large hemothorax
- Intra-abdominal hypertension/ compartment syndrome
The Debrief
- Recirculation is a phenomenon unique to VV-ECMO patients and can explain worsening hypoxia.
- Recirculation occurs when pump flow is diverted away from the RA/ TV and directly into the drainage cannula.
- Practically speaking, an increasing SpreO2 combined with a decreasing SaO2 may indicate clinically relevant recirculation is ongoing. A SpreO2 ≥ SaO2 cannot be explained without a recirculation event.
- Recirculation may be positional, or may be related to changing pressures in the chest or abdomen.
- Consider cannula-related and non-cannula causes for recirculation.
- Recirculation forms one of the logistical ceilings for escalating flow. When recirculation occurs, further increases in circuit flow likely will not improve oxygen delivery.
References
- Abrams, D. Brodie, D. Identification and management of recirculation in venovenous ECMO. 2015. Elso.org.
- Abrams D, Bacchetta M, Brodie D. Recirculation in venovenous extracorporeal membrane oxygenation. ASAIO J. 2015 Mar-Apr;61(2):115-21. doi: 10.1097/MAT.0000000000000179. PMID: 25423117.
- Patel B, Arcaro M, Chatterjee S. Bedside troubleshooting during venovenous extracorporeal membrane oxygenation (ECMO). J Thorac Dis. 2019;11(Suppl 14):S1698-S1707. doi:10.21037/jtd.2019.04.81
- Xie A, Yan TD, Forrest P. Recirculation in venovenous extracorporeal membrane oxygenation. J Crit Care. 2016 Dec;36:107-110. doi: 10.1016/j.jcrc.2016.05.027. Epub 2016 Jun 6. PMID: 27546757.