Extracorporeal membrane oxygenation (ECMO) is a therapy for respiratory and cardiac failure, providing support until organ recovery or destination therapy.1 The traditional modes of support are venovenous (VV), where blood is drained and reinfused via venous cannula(s), and venoarterial (VA), where blood is drained via a venous cannula and reinfused via an arterial cannula. In most cases, the VV mode is used for respiratory support and the VA mode for cardiorespiratory support. VV and VA modes are used in more than 95% of ECMO runs.2,3
Additional “hybrid” modes have been utilized when VV or VA modes do not provide sufficient support. One hybrid mode is veno-venoarterial (VVA), which is a circuit with a single venous drainage cannula and both venous and arterial reinfusion cannulas. The venous reinfusion cannula is added to increase the mixed venous oxygen concentration. This in turn increases the oxygen content ejected from the heart and the oxygen delivery to the head and upper body. This increased oxygen delivery may have the theoretical advantage of reducing neurological complications.4
The VVA mode of support is uncommon, but has been reported in case reports and a small case series.5–8 Most describe the use of this mode as a rescue for patients already on ECMO who develop worsening hypoxemia on VA support or cardiac dysfunction on VV support. It has also been described for North-South syndrome, which is a condition that can occur during VA ECMO with a femoral cannulation. In this scenario, the heart function is poor and the retrograde ECMO flow provides adequate oxygen delivery to the coronary and cerebral circulations. However, when the heart function recovers, it ejects blood against the retrograde flow from the ECMO circuit, creating a mixing point somewhere distally along the aorta. If the lung function is poor, the blood ejected from the heart is poorly oxygenated, and subsequently delivered to the head and upper body. This demarcation of flow between the heart and the circuit is physically observable with the head appearing blue and the lower body appearing pink.
The purpose of this study was to better characterize the population of patients supported with VVA ECMO, to understand when this mode is used, its impact on survival, and whether there was any reduction of complications, in particular neurological complications.
This HIPAA-compliant study was approved by the University of Michigan Institutional Review Board (HUM 00038817). The medical records of all patients at the University of Michigan Health System in Ann Arbor, Michigan who had a run of ECMO with a VVA configuration for any part of the run, between January 2000 and December 2014, were retrospectively reviewed. Collected data included the following: patient demographics (age, gender, past medical history, whether the patient was transferred), pre-ECMO course (ventilator days, therapy support adjuncts, last arterial blood gas before cannulation, whether a cardiac arrest occurred before cannulation), cannula configuration, ECMO course (duration, all ECMO modes employed, duration of VVA mode), and outcomes (lung recovery, survival to discharge, discharge location, hospital and ICU length of stay, complications).
Statistical analysis was performed to describe this population. Summary statistics were performed for each variable. Continuous variables were assessed for normalcy, and when normally distributed, were summarized by mean and standard deviation, otherwise were assessed with median and interquartile ranges. Categorical variables were summarized by frequency of observation.
There were 30 patients with 31 VVA runs between the years of 2000 and 2014. Adult patients composed 74% (23/31) of the cohort whereas pediatric patients were 26% (8/31).
There were 23 adult patients with 23 VVA runs. These patients were on average 40.4 years old; this population had an age range from 18 to 59 years old. Patients were 61% male (14/23). First, admission diagnosis was most commonly cardiac (10/23, 43%), which included elective cardiac surgery, congestive heart failure exacerbation, acute myocardial infarction, or heart transplant. Second most common were respiratory diagnoses (6/23, 26%) that included chronic obstructive pulmonary disease exacerbation or acute respiratory distress syndrome secondary to a pulmonary insult. Third were infectious diagnoses (3/23, 13%), which included sepsis from nonpulmonary sources that progressed to cardiopulmonary failure. Other diagnoses included elective gynecologic surgery and trauma. General patient characteristics are summarized in Table 1.
Adult patients were studied before extracorporeal life support. They were mechanically ventilated a median of 2.1 (1.0, 4.3) days before ECMO. Support adjuncts were used in all patients and included vasopressors (23/23, 100%), inhaled nitric oxide (6/23, 26%), and intra-aortic balloon pump (4/23, 17%). The last arterial blood gas before cannulation had an average pH of 7.24 ± 0.12, pCO2 46.5 (37.3, 63.5) mm Hg, pO2 55.0 (48.3, 77.3) mm Hg, bicarbonate 21.7 ± 8.5, and SpO2 81.8 ± 13.1%. The median PaO2/FiO2 ratio was 55 (48, 80). Nine patients (39%) had a cardiac arrest before being placed on ECMO support.
Adult patients were supported with ECMO for a median of 141 (97, 253) hours. Veno-venoarterial ECMO support time was a median of 110 (63, 179) hours. The majority (13/23, 57%) was first cannulated with a traditional ECMO mode: 7 (31%) with VA and 6 (26%) with VV. Ten (43%) were initially cannulated with a VVA approach. The VV supported patients were changed to VVA for cardiac failure (6/13, 46%), which presented as refractory shock with increasing vasopressor requirements and increasing lactate concentrations. The VA-supported patients were changed to VVA for North-South syndrome (5/13, 38%) or for generalized worsening hypoxia (2/13, 15%) despite circuit optimization with blood transfusion or flow adjustment. Patients spent 97% (64, 100) of their ECMO run in the VVA configuration.
Multiple cannulation strategies were employed. The majority of patients (18/32, 78%) had a lower body venous drainage, lower body arterial reinfusion, and an upper body venous reinfusion cannula. Other observed configurations included using a double lumen internal jugular venous cannula with a lower body arterial reinfusion (2/23, 9%); lower body venous drainage with upper body arterial and venous reinfusion (1/23, 4%); lower body venous drainage with lower body arterial and venous reinfusion (1/23, 4%); and double lumen internal jugular venous cannula with upper body arterial reinfusion (1/23, 4%). For those patients with lower body venous and arterial cannulas, a posterior tibial reinfusion cannula was used 47% time. Cannulation configuration strategies are illustrated in Figure 1.
Outcomes of the adult patients were evaluated and summarized in Table 2. It was found that 48% (11/23) had lung recovery and were able to be weaned off ECMO. Of these patients, 82% (9/11) survived to discharge; the overall survival of adult VVA patients was 39%. Surviving patients were discharged to home (6/9, 67%), extended-care facilities (2/9, 22%), or outside hospital intensive care units (1/9, 11%). For this cohort, the median hospital length of stay was 30 (18, 41) days, ICU length of stay was 19 (15, 31) days, and ventilator-free days in 60 days was 4 (0, 49) days.
All complications of the adult patients are shown in Table 3. Neurologic complications were observed in three patients (13%). Of these patients, 1 had seizures, 1 had a stroke, and 1 had clinically-determined brain death and stroke. Limb ischemia complicated five adult patients (22%). One patient required amputation because of ischemia; however, the other limbs were salvaged by placement of a tibial reperfusion catheter.
There were seven pediatric patients with eight VVA runs. These patients were on average 13.0 years old; this cohort had an age range from 11 to 17 years old. Patients were 43% male (3/7). Admission diagnosis was most commonly pulmonary (4/7, 57%), which included influenza or pneumonia; other diagnoses were right heart failure secondary to pulmonary embolism, classified as cardiac (1/7, 14%); septic knee joint, classified as infectious (1/7, 14%); and a trauma (1/7, 14%). Pediatric patient characteristics are summarized in Table 1.
Pediatric patients were studied before extracorporeal life support. They were mechanically ventilated a median of 1.1 (0.9, 2.1) days before ECMO. Support adjuncts were used in all patients and included vasopressors (8/8, 100%) and inhaled nitric oxide (3/8, 38%). The last arterial blood gas before cannulation had an average pH of 7.19 ± 0.08, pCO2 55.0 (46.8, 65.0) mm Hg, pO2 53.5 (42.0, 63) mm Hg, bicarbonate 20.4 ± 6.1, and SpO2 77.8 ± 15.9%. The median PaO2/FiO2 ratio was 54 (42, 63). Three patients (38%) had a cardiac arrest before being placed on ECMO support.
Pediatric patients were supported with ECMO for a median of 258 (168, 419) hours. Veno-venoarterial extracorporeal membrane oxygenation support time was a median of 131 (98, 161) hours. The majority (7/8, 87.5%) was first cannulated with a traditional ECMO mode: 5 (62.5%) with VA and 2 (25%) with VV. One (12.5%) was initially cannulated with a VVA approach, and this was the second run for a patient whose first run was converted to VVA. The patients supported with VV ECMO were changed to the VVA mode for cardiac failure (2/7, 29%). Patients originally supported with VA mode changed to VVA for North-South syndrome (3/7, 42%) or generalized worsening hypoxia (2/7, 29%). Patients spent 46% (37, 74) of their ECMO runs in the VVA configuration.
Two cannulation strategies were employed. The majority of patients (6/8, 75%) had a lower body venous drainage, lower body arterial reinfusion, and an upper body venous reinfusion cannula. The other observed configuration was unusual for the VVA mode in that the carotid artery was ligated. It consisted of upper body venous drainage, upper body arterial reinfusion, and lower body venous reinfusion (2/8, 25%). Pediatric cannulation configuration strategies are illustrated in Figure 2. For those patients with lower body venous and arterial cannulas, a posterior tibial reinfusion cannula was used 67% time.
Outcomes of the pediatric patients are summarized in Table 2. Of these patients, 75% (6/8 runs) had lung recovery and are able to be weaned off ECMO. Patient survival to discharge was 57% (5/7). Surviving patients were discharged to home (1/5, 20%), extended care facilities (3/5, 60%), or outside hospital intensive care units (1/5, 20%). For this cohort, the median hospital length of stay was 31 days (24, 42), ICU length of stay was 24 days (21, 29), and ventilator-free days in 60 days was 34 (0, 42) days.
All complications of pediatric patients are shown in Table 3. Neurologic complications were reported in two patients (29%). Of these patients, one had seizures and one had a large intracranial hemorrhage. Limb ischemia complicated two pediatric patients (29%), which was treated with placement of a tibial reperfusion cannula.
These data demonstrate that patients supported with a VVA mode are a small, heterogeneous cohort. We had only 30 patients in the study, collected over a 14 year period, even though our institution is a high-volume center having supported more than 2,000 patients on ECMO.10 The vast majority of our patients received femoral VA support with internal jugular reinfusion to improve the oxygen delivery to the upper body. With rare exceptions, this strategy was used to avoid carotid cannulation when cardiopulmonary support was required and upper body perfusion was inadequate. As such, it was hypothesized that neurologic complications would be minimized. However, we did not find a survival advantage or diminution in complications in this study population.
This cohort has some similarities to other ECMO populations. The adult subjects most closely matched cardiac-supported patients placed on VA ECMO. Although VVA ages were younger, their comorbid status, frequency of cardiac arrest before cannulation, and ventilator time before cannulation were comparable.10–12 Survival of only 39% also mirrored the reported survival of cardiac-supported VA ECMO patients (40%).11,12 This survival was consistent with the survival of adult VA patients at our institution over a similar period (40%) but lower than the adult VV supported patients (48–55%).10 The pediatric cohort had minimal co-morbidities before ECMO and survival was 71%. This survival was intermediary between pediatric cardiac and respiratory supported patients.10
The overall rate and types of complications for adults and pediatrics were not appreciably different than other reports or our internal experience with traditional ECMO modes (VV, VA) during a similar time period.10 Neurological complications in adults, observed in 13% of our adult runs, was consistent with other series.3,8,12 This suggests that neurologic complications during ECMO are multifactorial and likely not solely related to upper body oxygen delivery. Because of the retrospective nature of this study, we are unable to differentiate the causation of the neurocognitive damage; it could be due to the period of poor oxygen delivery or secondary to the poor perfusion necessitating ECMO support.
The pediatric cohort had a neurologic complication rate of 29%, which was higher than expected based on cannulation strategy. In a recent review of the ELSO database assessing children less than 18 years of age, the risk of neurologic complications for carotid cannulation was 23%, compared with 17% for aortic cannulation, and 15% for femoral cannulation.13 Interestingly, the two patients in our cohort who had neurologic complications did not have carotid cannulation. This difference may be due in part to small sample size and selection bias of our pediatric cohort. Additionally, the upper body venous reinfusion catheter may create cerebral venous hypertension that could contribute to neurologic injury.14 For children, our data suggests that because the VVA mode did not reduce neurologic complications, VA cannulation using the neck vessels should still be considered in an unstable patient requiring cardiopulmonary support.
A potential drawback of the VVA mode is the risk of leg ischemia, because the majority of patients had reinfusion through the femoral artery. The rates of leg ischemia associated with femoral artery cannulation for VA ECMO range between 20% and 52%.15–18 Our overall rates of limb complications were 25% and 22% for adults and pediatric patients, respectively. Institutionally, we introduced posterior tibial reinfusion cannulas in 2005, after which our rates of leg ischemia decreased substantially19: no adult or pediatric patients in this cohort had limb complications if a posterior tibial catheter was used prophylactically or within 6 hours of ECMO initiation.
Since survival is not improved nor complications reduced with a VVA mode, one may ask when should this mode be used? We propose there are a few clinical scenarios in which it should be considered primarily in adult patients. First, it is an effective solution to the North-South syndrome. This occurs in the context of VA support with femoral artery reinfusion when the oxygenated blood, pumped retrograde into the aorta, does not perfuse the upper body. This can be confirmed by hypoxemia on a blood gas taken from the right radial artery, which is a surrogate marker for cerebral and coronary circulations. By adding a second venous infusion catheter, oxygen delivery to the upper body is improved by improving the mixed venous oxygen concentration.6,7 This may be even more important with patients who are older and have cerebral vascular or coronary disease.4 A second scenario is with patients who develop respiratory dysfunction, are placed on VV ECMO, but later develop cardiac dysfunction necessitating VA support. Clinically, this is observed as low mean arterial blood pressure with increasing vasopressor or inotropic requirements or decreasing mixed venous saturations (SvO2) despite stable ECMO circuit flows. It also can be confirmed with echocardiography. In this situation, the original cannula(s) are left in place and an arterial infusion catheter is added, creating a hybrid VVA circuit.5 A final scenario, although not observed in this study cohort, is for a patient with cardiopulmonary dysfunction and intrinsic pulmonary disease. The infusion of oxygenated blood into the pulmonary vasculature might enhance pulmonary artery vasodilation and help off load right heart dysfunction.
There are a few technical concerns that should be considered when the choice to convert a patient to the hybrid VVA ECMO mode has been made. When choosing cannulation sites, we leave the existing cannulas in place and add the cannula necessary to complete the VVA configuration. We prefer the right internal jugular vein when a venous reinfusion needs to be added because of ease of placement and the oxygen delivery to the coronary and cerebral circulations is higher. When an arterial cannula is added in adults, we favor a femoral site over the carotid to reduce the stroke risk. In pediatrics, there are recent data13 demonstrating acceptable stroke rates with carotid cannulation. In our cohort, however, the two pediatrics patients who went from VV to VVA had femoral arterial cannulas placed. After cannulation, the venous and arterial reinfusion limbs are connected with a “Y” connector. A Hoffman clamp is placed on the venous reinfusion limb to decrease stealing of blood from the arterial reinfusion, as pulmonary vascular resistance is less than systemic vascular resistance. Flows are initially begun with 25% through the venous reinfusion and 75% through the arterial reinfusion, but fine adjustments are made until acceptable blood pressure and mixed venous saturations are attained and the clinical situation prompting the change to VVA is resolved.
Management of VVA ECMO was similar to other modes of ECMO at our institution. Patient hypoxemia was monitored both with an inline SvO2 detector in the venous drainage limb of the circuit and through pulse oximetry. In adults, systemic mean arterial pressures were targeted to be greater than 60 mm Hg; increased flow through arterial reinfusion limb or vasopressors were used to achieve this goal. The circuits were anticoagulated using unfractionated heparin, with a target of ACT between 210–230 or anti-Xa level between 0.3 and 0.5, unless patient had a bleeding complication, and then the anticoagulation target was reduced or held. Other ECMO best practices, such as monitoring daily hemolysis laboratories and inspecting the circuit for thrombus, were performed. Weaning from VVA ECMO was dependent on organ recovery. If the cardiopulmonary injury had resolved, then flow and sweep were systematically and step-wised decreased until appropriate levels were achieved to consider trialing off. If lung recovery lagged behind cardiac recovery, the circuit was converted from VVA to VV ECMO.
Because circuits with multiple venous and arterial cannulas are infrequently used, the nomenclature is often confusing. In this articles, we used the nomenclature established by the Extracorporeal Life Support Organization (ELSO),2 although other variations exist.5 As per the ELSO definition, VVA is a mode of cannulation with a single drainage cannula and both a venous and arterial reinfusion cannulas. This nomenclature is used even if the patient started in a VV or VA configuration and then converted to a VVA mode. Other modes that may be confused with VVA are VA + V, which denotes VA support with an additional venous drainage cannula, or VV − AV, which represents a configuration change from VV to VA ECMO support.
There are several important limitations to this study. It was a single-center retrospective study that spanned 14 years. During this period, many practice patterns have changed. The number of patients is small, particularly for a heterogeneous cohort that limited us to a descriptive study. The small number of patients coupled with the absence of a matched control group did not allow for robust statistical analysis and limited the conclusions.
Veno-venoarterial extracorporeal membrane oxygenation is uncommon. Although theoretically appealing, we observed neither survival advantage nor complication reduction using the VVA mode of support. However, it does have value for unique clinical situations, and should be part of the clinician’s armamentarium of salvage therapies when conventional ECMO modes do not meet patient support needs. Specifically, adult patients may benefit most from this strategy given the association of carotid cannulation and increased stroke rate.
1. Bartlett RH: Extracorporeal life support for cardiopulmonary failure. Curr Probl Surg 1990.27: 621–705.
2. Paden ML, Rycus PT, Thiagarajan RR: Update and outcomes in extracorporeal life support. Semin Perinatol 2014.38: 65–70.
3. Brogan TV, Thiagarajan RR, Rycus PT, Bartlett RH, Bratton SL: Extracorporeal membrane oxygenation in adults with severe respiratory failure: A multi-center database. Intensive Care Med 2009.35: 2105–2114.
4. Abrams D, Combes A, Brodie D: Extracorporeal membrane oxygenation in cardiopulmonary disease in adults. J Am Coll Cardiol 2014.63: 2769–2778.
5. Biscotti M, Lee A, Basner RC, et al: Hybrid configurations via percutaneous access for extracorporeal membrane oxygenation: A single-center experience. ASAIO J 2014.60: 635–642.
6. Moisan M, Lafargue M, Calderon J, Oses P, Ouattara A: Pulmonary alveolar proteinosis requiring “hybrid” extracorporeal life support, and complicated by acute necrotizing pneumonia. Ann Fr Anesth Reanim 2013.32: e71–e75.
7. Lingadevaru H, Romano MA, Fauman K, Cooley E, Annich GM, Cornell TT: Challenges during repeat extracorporeal life support in a patient with pulmonary alveolar proteinosis. ASAIO J 2011.57: 473–474.
8. Stöhr F, Emmert MY, Lachat ML, et al: Extracorporeal membrane oxygenation for acute respiratory distress syndrome: Is the configuration mode an important predictor for the outcome? Interact Cardiovasc Thorac Surg 2011.12: 676–680.
9. Knaus WA, Draper EA, Wagner DP, Zimmerman JE: APACHE II: A severity of disease classification system. Crit Care Med 1985.13: 818–829.
10. Gray BW, Haft JW, Hirsch JC, Annich GM, Hirschl RB, Bartlett RH: Extracorporeal life support: Experience with 2,000 patients. ASAIO J 2015.61: 2–7.
11. Xie A, Phan K, Tsai YC, Yan TD, Forrest P: Venoarterial extracorporeal membrane oxygenation for cardiogenic shock and cardiac arrest: A meta-analysis. J Cardiothorac Vasc Anesth 2015.29: 637–645.
12. Truby L, Mundy L, Kalesan B, et al: Contemporary outcomes of venoarterial extracorporeal membrane oxygenation for refractory cardiogenic shock at a large tertiary care center. ASAIO J 2015.61: 403–409.
13. Teele SA, Salvin JW, Barrett CS, et al: The association of carotid artery cannulation and neurologic injury in pediatric patients supported with venoarterial extracorporeal membrane oxygenation*. Pediatr Crit Care Med 2014.15: 355–361.
14. Woda RP, Miner ME, McCandless C, McSweeney TD: The effect of right internal jugular vein cannulation on intracranial pressure. J Neurosurg Anesthesiol 1996.8: 286–292.
15. Gander JW, Fisher JC, Reichstein AR, et al: Limb ischemia after common femoral artery cannulation for venoarterial extracorporeal membrane oxygenation: An unresolved problem. J Pediatr Surg 2010.45: 2136–2140.
16. Kimura N, Kawahito K, Ito S, et al: Perfusion through the dorsalis pedis artery for acute limb ischemia secondary to an occlusive arterial cannula during percutaneous cardiopulmonary support. J Artif Organs 2005.8: 206–209.
17. Ko WJ, Lin CY, Chen RJ, Wang SS, Lin FY, Chen YS: Extracorporeal membrane oxygenation support for adult postcardiotomy cardiogenic shock. Ann Thorac Surg 2002.73: 538–545.
18. Smedira NG, Moazami N, Golding CM, et al: Clinical experience with 202 adults receiving extracorporeal membrane oxygenation for cardiac failure: Survival at five years. J Thorac Cardiovasc Surg 2001.122: 92–102.
19. Spurlock DJ, Toomasian JM, Romano MA, Cooley E, Bartlett RH, Haft JW: A simple technique to prevent limb ischemia during veno-arterial ECMO using the femoral artery: The posterior tibial approach. Perfusion 2012.27: 141–145.
ECMO; hybrid; veno-venoarterial; VVA; salvage