Abstract
Introduction:
Different cannulation approaches existed for veno-venous extracorporeal membrane oxygenation (VV ECMO). We aimed to compare the atrio-femoral (AF) and femoro-atrial (FA) configuration in terms of their flow efficiency and influence on patient outcome.
Method:
This was a single-centre, retrospective case control study. Adult patients admitted to the Intensive Care Unit and required VV ECMO service at Tuen Mun Hospital, Hong Kong, from June 2015 to January 2020 were included. Data were collected from our ECMO database for comparison.
Results:
Between June 2015 and January 2020, eight patients received AF configuration and 19 patients received FA configuration. The maximum achieved flow in the AF group was significantly higher than that in the FA group (4.08 ± 0.57 L/min vs. 3.52 ± 0.58 L/min, p = 0.03). The fluid balance in first 3 days of ECMO was significantly lower in the AF group compared to that in the FA group (1.16 ± 2.71 L vs. 3.46 ± 1.97 L, p = 0.02). As well, the chance for successful awake ECMO was statistically higher in the AF group (p = 0.048).
Conclusion:
Atrio-femoral configuration in VV ECMO was associated with a higher maximum achieved ECMO flow, less fluid gain in first 3 days of ECMO and more successful awake ECMO.
Introduction
Extracorporeal membrane oxygenation (ECMO) has been used for patients with respiratory failure for over three decades. 1 Through the history, different cannulation approaches for veno-venous (VV) ECMO have been described. Both one dual-lumen cannulation and two single-lumen cannulation approaches exist. For single-lumen cannulation, the most common ways would either be draining blood from right atrium (RA) via a jugular cannula and returned to inferior vena cava (IVC) via a femoral cannula, which is called atrio-femoral (AF) approach, or from the IVC via a femoral cannula and return to RA via a right jugular cannula, which is called femoro-atrial (FA) approach. 2 In the early days of VV ECMO, both approaches were employed until the study done by Rich et al. who compared AF and FA flow in adult VV ECMO. 3 In the study, both the right internal jugular vein and femoral vein of patients were cannulated with a single-lumen single-stage cannula which were connected to an ECMO circuit with a modified bridge to allow switch between AF and FA flow direction. The authors found that FA approach provided a higher maximum flow, higher pulmonary artery SvO2 and required a lower flow to maintain the same SvO2 than AF approach did. Since then FA approach has been widely employed and recommended by majority of the ECMO centres. 4 However, argument about this study existed since all study patients were paralyzed and subjected to pressure-controlled inverse-ratio ventilation to maintain pulmonary recruitment which may cause congestion of venous blood in the IVC, and a certain degree of RA collapse and hence drainage insufficiency from RA. In addition, the study was performed on the first day of ECMO support, thus weaning and fluid withdrawal were not attempted. 2 In a study done by a Swedish ECMO centre, the author suggested that, in AF approach, since drainage occurs in RA which is a cavity with a large capacitance, adequate flow can be maintained longer despite ambitious fluid withdrawal. As well, by using a multistage cannula, recirculation fraction in AF approach can be significantly reduced as compared to a single-stage cannula, thus improving effective ECMO flow. 5 In our centre, we performed both atrio-femoral and femoro-atrial cannulation for our patients who required VV ECMO.
The aim of this study was to compare atrio-femoral (AF) and femoro-atrial (FA) configuration in terms of their flow efficiency and influence on patient outcome.
Method
This was a single-centre, retrospective case control study on differences between atrio-femoral (AF) and femoro-atrial (FA) configuration in VV ECMO.
Study population
Adult patients admitted to the Intensive Care Unit (ICU) and required VV ECMO service at Tuen Mun Hospital, Hong Kong, from June 2015 to January 2020 were included. Exclusion criteria were patients who were less than 18 years old, who were subsequently transferred to other ECMO centre due to clinical needs, who were cannulated with other types of configuration, for example, femoro-femoral and dual-lumen bicaval cannulation, who subsequently developed cardiogenic shock and required conversion to venoarterial (VA) ECMO, and who had massive haemorrhage requiring massive transfusion during ECMO.
Criteria for VV ECMO
Based on the local guideline by the Hospital Authority Coordinating Committee in Intensive Care in Hong Kong, patients in our unit with severe ARDS were indicated for VV ECMO if, after optimal ventilation with positive end-expiratory pressure (PEEP) >10 and plateau pressure > 30 cmH2O, they had a P/F ratio < 60 mmHg, a P/F ratio < 80 mmHg with a Murray score of 3 to 4, or a P/F ratio < 100 mmHg with pCO2 > 80 mmHg for over 1 hour.
Patients with significant pre-existing comorbidity, poor performance status, severe pulmonary hypertension not for transplant and irreversible pulmonary pathology were considered contra-indicated for VV ECMO.
Cannulation
In our unit, we used Cardiohelp (Maquet Cardiopulmonary AG, Rastatt, Germany) and Rotaflow (Maquet Cardiopulmonary AG, Rastatt, Germany) consoles. All cannulations were done by attending intensive care physicians who were trained in ECMO. Choice of configuration was up to the preference of the attending physician.
For AF configuration, a 23 Fr/38cm multistage Maquet HLS (Maquet) drainage cannula was inserted in the right internal jugular vein with its tip in the RA as guided by transthoracic echocardiography. A 19 to 21 Fr/23cm single-stage Maquet HLS (Maquet) return cannula was inserted in the femoral vein with its tip at the iliac vein.
For FA configuration, a 21 to 25 Fr/55cm multistage Maquet HLS (Maquet) drainage cannula was inserted in the femoral vein with its tip at the IVC/RA junction as guided by transthoracic echocardiography. A 15 to 21 Fr/23cm single-stage Maquet HLS (Maquet) return cannula was inserted in the right internal jugular vein with its tip at the superior vena cava (SVC) or RA.
Patient management on ECMO
After cannulation, similar to the practice of most ECMO centres, 6 ECMO blood flow was adjusted to reduce the fraction of inspired oxygen (FiO2) of mechanical ventilator to less than 0.6 and maintain arterial saturation (SaO2) more than 88%. Sweep gas flow was adjusted to achieve a normal pH, tidal volume less than 6 ml/kg ideal body weight and plateau pressure less than 25 cmH2O. In our practice, synchronised intermittent mandatory ventilation (SIMV) with pressure control (PC) and pressure support (PS) ventilatory mode was used for all of our patients on ECMO. PEEP was maintained above 8 to 10 cmH2O. Anticoagulation with unfractionated heparin was maintained to keep activated partial thromboplastin time (APTT) 1.5 to 2 times the normal value of patients.
Negative fluid balance and minimisation of sedation were the general goal for our ECMO patients. If ECMO flow was insufficient to maintain oxygenation and lung protective ventilation, fluid boluses, sedatives and/or neuromuscular blockers could be given to reduce oxygen consumption and enhance efficiency of ECMO support. Change of configuration to high flow VV-V ECMO would be considered if the ECMO flow was insufficient or tube chattering issue was refractory to the above measures.
Awake ECMO, that is, ECMO without intubation and mechanical ventilation, would be considered if patient was alert and cooperative, had low work of breathing, was haemodynamically stable and was not expected to be weaned from ECMO in the coming 48 hours. If a patient was extubated and did not require reintubation in order to facilitate ECMO support, it was defined as successful awake ECMO in this study.
For ECMO weaning, after improvement of native lung function, ECMO flow was gradually reduced to 2 to 2.5 L/min and the sweep gas flow was then stopped for 4 hours. ECMO was weaned if blood gases remained stable.
Data collection
Data including patient demographics, causes of ARDS, acute physiology and chronic health evaluation (APACHE) score, days on ECMO, need of renal replacement therapy (RRT) in the first 3 days of ECMO, number of awake ECMO, ICU length of stay and survival to ICU discharge were collected from our ECMO database.
Outcome variables
Outcome variables including the maximum achieved ECMO flow during the whole ECMO support, fluid balance in first 3 days of ECMO, need of paralysis to facilitate ECMO support, need of conversion to high flow VV-V configuration and successfulness of awake ECMO were collected from our ECMO database for comparison.
Statistical analysis
IBM SPSS Statistics Version 26.0 (IBM Corp., Armonk, NY) was used for statistical analysis. Numerical data were expressed as mean ± standard deviation and compared with Mann-Whitney U test. Categorical data were analysed with Fisher’s exact test. A p value <0.05 was considered significant.
Ethics
This study was approved by New Territories West Cluster Research Ethics Committee (NTWC REC).
Results
Between June 2015 and January 2020, a total of 36 patients received VV ECMO in our unit. After exclusion of nine patients, 27 patients were included in the analysis. Eight of them received VV ECMO with AF configuration while 19 of them had FA configuration as shown in Figure 1. The characteristics of the study population were shown in Table 1. The parameters were similar except that the AF group carried more patients with combined viral and bacterial infection.
Study population of patients who required venovenous (VV) ECMO in our unit from 6/2015 to 1/2020.
Characteristics of the study population.
ARDS: acute respiratory distress syndrome; APACHE: acute physiology and chronic health evaluation; ICU: intensive care unit; RRT: renal replacement therapy; ECMO: extracorporeal membrane oxygenation.
Data are number (%) or median (Interquartile range).
The maximum achieved flow in the AF group was significantly higher than that in the FA group (4.08 ± 0.57 L/min vs. 3.52 ± 0.58 L/min, p = 0.03). In addition, the fluid balance in first 3 days of ECMO was significantly lower in the AF group compared to that in the FA group (1.16 ± 2.71 L vs. 3.46 ± 1.97 L, p = 0.02). Five out of eight patients in the AF group fuifilled criteria for and underwent successful awake ECMO. However, three out of four patients who underwent awake ECMO in the FA group required reintubation due to drainage insufficiency, thus considered as failure of awake ECMO. Thus, chance for successful awake ECMO was statistically higher in the AF group (p = 0.048).
On the other hand, the need of paralysis to facilitate ECMO support and conversion to high flow VV-V ECMO did not differ between the groups (p = 0.68 and p = 0.29, respectively).
Discussion
Our study showed that AF configuration provided a higher maximum achieved ECMO flow. This was in contrast to the study done by Rich et al. There might be several reasons for this. Firstly, RA is a cavity with a larger volume and capacitance than the IVC. This laid the foundation of a smooth venous drainage by the jugular cannula. Furthermore, we used a multistage drainage cannula which differed from the single-stage cannula used by Rich et al., allowing a lower pressure drop at the same ECMO flow. In addition, commonly used ventilatory mode, that is, SIMV with PC/PS mode, was used for our ECMO patients. The magnitude of positive airway pressure was likely to be lower than that of pressure-controlled inverse-ratio ventilation, hence the effect on the volume and capacitance of right atrium was lower.
A higher achievable ECMO flow can lead to a smoother ECMO management. This was shown in our result that patients in the AF group required less fluid in the first 3 days of ECMO in order to keep a constant venous drainage. In some of our patients in the AF group, even fluid withdrawal could be safely implemented to achieve a negative fluid balance within first 3 days of ECMO support. Furthermore, as previously shown, a positive fluid balance in first few days of ECMO is associated with a higher hospital and 90-day mortality.7,8 Though such result could not be demonstrated in our results, the use of AF configuration may have an influence on mortality of ECMO patients. A study with a larger sample size is needed to evaluate such potential. In addition, drainage insufficiency with frequent line chattering is known to induce cavitation, haemolysis 9 and potentially damage venous vessel wall. A smooth ECMO flow could reduce such complications.
There are many possible advantages of awake ECMO. Apart from having a better ventilation-perfusion matching, ventilator-induced lung injury, ventilator induced diaphragm dysfunction and ventilator associated pneumonia can be avoided. The absence of endotracheal tube also means less use of sedatives and analgesics and hence lower incidence of delirium 10 and patients can regain their autonomy and take active part in daily care. In our study, more successful awake ECMO could be achieved in the AF group compared to the FA group. We attributed the reason to the smoother ECMO drainage in AF configuration. From our experience, with a smoother ECMO flow, less sedatives and analgesics are needed to facilitate ECMO support, patients could therefore awaken and be weaned from ventilator more quickly. In addition, ECMO flow can be maintained even if patients take deep breath and cough during spontaneous breathing. Furthermore, the patient can enjoy a certain degree of mobilisation and rehabilitation like passive cycling without influencing the ECMO flow.
There were several limitations in our study. We were not able to measure recirculation fraction in our ECMO patients since ultrasound dilution technology is not available in our unit. AF configuration have been criticised to have a higher recirculation fraction limiting the effective blood flow. Whenever we encountered recirculation with inadequate arterial oxygen saturation, we increased the ECMO flow first despite the fact that a portion of flow belonged to recirculating blood. We also separated the cannula further by retracting the return cannula in both AF and FA configuration. Another option was transfusion of red cells in order to increase the oxygen carrying capacity. Lactate level was monitored in our ECMO patients to watch out for inadequate oxygen delivery. From our experience, all our patients in the AF group had adequate oxygenation, while four patients in the FA group did not receive adequate blood flow and required conversion to high-flow VV-V configuration. This underlined the importance of a higher margin of flow adequacy provided by AF configuration.
Furthermore, confounding existed in our study as ECMO flow in the two groups might be affected by cannula of different lengths and sizes which may not allow direct comparison. The 55 cm long Maquet HLS (Maquet) multistage cannula in the FA group has a higher flow resistance compared to the 38 cm long Maquet HLS (Maquet) multistage cannula in the AF group, as explained by the Hagen-Poiseuille equation. However, based on the evidence that most of the flow occur at the most proximal side holes, 11 that is, 20 cm from tip for the 55 cm cannula and 10 cm from tip for the 38 cm cannula, the difference in flow resistance between the two cannulas is lower than expected. As the 55 cm long cannula is the most widely used cannula nowadays while the 38 cm long cannula is the only short multistage cannula available on the market, we believed that comparison of the groups with the most widely used cannula would be more realistic rather than pursuing absolute equality between the groups. We chose the size of drainage cannula depending on patients’ body size and the flow goals, 23 Fr drainage cannula, which was our most commonly used size in both the AF and FA groups, was also the average size cannula used in other ECMO centres. 12
Effects of patient’s lung status and ventilator settings on ECMO flow were other issues which were hard to address as patient’s lung compliance changed over time. We tried to compare the maximum flow that could be achieved during the ECMO support rather than the mean flow. In addition, the effects were reduced by having groups with similar disease severity and standardised ventilator setting targeting lung protective ventilation for our patients.
Moreover, this study was not a randomised controlled trial. Choice of configuration depended on physician’s preference. As with all observational studies, our results were prone to bias. A prospective randomised controlled trial would be optimal to compare the two configurations.
Conclusion
Atrio-femoral configuration in VV ECMO was associated with a higher maximum achieved ECMO flow, less fluid gain in first 3 days of ECMO and more successful awake ECMO.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship and/or publication of this article.
