Monitoring a Veno-Venous ECMO: the Role of Novel Real-Time Continuous Monitoring Technology

Evolution of extracorporeal circulation technologies has produced a breakthrough in the care of patients with severe respiratory failure. The concept of massive extracorporeal support has spread out from cardiac surgery operating theatres and few specialized intensive care units and is an option that clinicians bear in mind whilst taking care of such patients.Veno-Venous ExtraCorporeal Membrane Oxygenation (VV-ECMO) is briefly an artificial Membrane Lung (ML), with its blood pump, placed in series with the failing Natural Lung (NL). Thus the ML can totally or partially take over the functions of the NL in both carbon dioxide removal and oxygen intake. The delivery of O₂ to the patient’s metabolism and the removal of CO₂ depend on a complex interaction between the ML, the NL and the metabolic status. The simple number of oxygen saturation (SpO₂) on the screen can be modified by various elements or by different situations, some of which can need a prompt response. ML, NL and the metabolic-hemodynamic pattern of the patient must be analyzed separately in order to understand what’s happening and what’s the correct management of the situation.
Monitoring the ML is essential to understand the level of performance of gas exchange; this is paramount in order to avoid abrupt failures of the ML, as a failing circuit may be diagnosed when performances are still acceptable and changed in a scheduled manner when staff, expertise and setting are available thus improving safety in a procedure that may be risky and unpredictable. The NL function in both gas exchange and mechanical properties must be evaluated in order to realize the possibility and timing of VV-ECMO weaning and removal. The timing of ECMO is as important and debated as its technical choices: if weaning is too early, ventilation may strain a convalescing but yet not entirely recovered lungs thus decreasing their improvement and maybe triggering long term morbidity; if the VV-ECMO circuit is left in function on a NL that can safely answer to the metabolic needs of our patients, we may expose them to unneeded risks and complications, especially of hemorrhagic and infectious nature. Monitoring the hemodynamic and metabolic parameters of our patient (Cardiac Output, V O₂, V CO₂), apart from the general use in critical care, is paramount to understand and achieve an equilibrium between the metabolic needs and the function of the artificial organ; this may be quite demanding in septic or extremely obese patients in which metabolic needs may need an absolutely performant oxygenation and decarboxylation. The ratio between Cardiac Output (CO) and extracorporeal Blood Flow (BF) is fundamental in understanding recirculation in the ML and difficulties in obtaining satisfactory BF.

Fig.1 A schematic view of a VV-ECMO with the ML and NL in series and the patients metabolism. Both lungs have a Dead Space with ventilation but no perfusion, a Shunt with perfusion but no ventilation and percentage of blood that has an Ideal V/Q (Ventilation/Perfusion ratio). These three components are to be assessed in order to evaluate the performance of the ML and the situation of the NL.
BF, Blood Flow; CO, Cardiac Output; V O₂ Oxygen consumption; V CO₂ Carbon Dioxide production

We may now focus to monitoring the ML, which is peculiar to VV-ECMO. The ML can be considered exactly as a form of NL, as its main functions are to add oxygen to the blood (V’ O2 ML) and to remove carbon dioxide (V’ CO₂ ML). These two functions must be checked separately as they may not always decay at the same time and we may be forced to change a ML for failure of one function while the other is still more than efficient.
Measuring the Oxygenating Power of the Membrane Lung
The amount of oxygen added by the ML can be calculated as follows:
V’ O₂ _ML = BF x ( Cont O_{2 OUT} – Cont O_{2 IN}) x 10
where V’ O_{2 ML} is measured in mL/min, Blood Flow (BF) is in L/min, Cont O_{2 OUT} (Oxygen content in the blood coming out of the ML) and Cont O_{2 IN} (Oxygen content in the blood entering the ML) in mL/dL. As blood content in oxygen is as follows:
Cont O₂ = P O₂ x 0,0031 + Hgb x S O₂ x 1,34
we realize that haemoglobin concentration (Hgb in g/dL) and its saturation in oxygen (S O₂ in %) are of paramount importance in the amount of oxygen carried by blood while partial oxygen pressure (P O₂ in mmHg), being multiplied by 0,0031 (the solubility coefficient), has a negligible role.
We can notice also that V’ O₂ _ML is directly connected to BF. Obviously it’s the portion of BF that enters in contact with gas flow (perfusion coupled with ventilation) that is important for oxygenation. The portion of Bf that is not ventilated, i.e. shunt, doesn’t uptake any oxygen; the higher the shunt ratio the lower the V’O2 ML. New ML have already a minor shunt ratio of around 10-20%, which then increases as debris and fibrin coat the capillaries in which the blood flows thus hindering the passage of oxygen.
Shunt can be calculated as
Shunt/BF = ( Cont O₂ _CAP – Cont O_{2 OUT} ) / (Cont O_{2 CAP} – Cont O_{2 IN} )
Where Cont O_{2 CAP} is calculated with the Cont O₂ formula using [ 100 – ( Hgb CO + Met Hgb ) ] as ( S O₂ ) and [ ( P_ATM-P_H2O) x FiO₂ – P CO₂ / RQ ] as P O₂ where Hgb CO is haemoglobin saturated in carbon monoxide, Met Hgb is Methemoglobin, P_ATM is atmosphere pressure, P_H2O is steam partial pressure, FiO₂ is inhaled O₂ fraction, PCO₂ is blood CO₂ partial pressure and RQ is the Respiratory Quotient.
Though intellectually intriguing, it is no surprise that shunt ratio is not widely used as a trouble shouting parameter due to its cumbersome calculation. The PaO2/FiO2 is clearly a much more user friendly parameter for the ML, considering also that we don’t have the Mean Airway Pressure (MAP) or Positive End Expiratory Pressure (PEEP) to take into account. Also the PaO₂/FiO₂ and Shunt/BF ratios are inversely proportional.
In brief the V’ O₂ ML is increased with:
• Increase of BF,
• Increase of FiO₂,
• Reduction of blood oxygen saturation at the ML inlet,
• Reduction of the recirculation ratio in the ML.
On the contrary V’ O₂ ML is reduced with:
• increase of the shunt ratio (as we can see in normal ML after a long period of use).

Measuring the Decarboxylation Power of the Membrane Lung

VV-ECMO has an enormous capacity of clearing the patient’s blood of CO₂. The V’ CO₂ ML can be easily measured with a CO₂ analyser on the gas outlet of the ML which gives the P CO₂, the carbon dioxide partial pressure, of the air coming out of the ML. Having measured the Gas Flow (GF), we can obtain the V’ CO₂ ML as follows:
V’ CO₂ ML = P CO₂ x GF x (1000/760)
where GF is measured in L/min, P CO₂ in mmHg and V’ CO₂ ML in ml/min.
The capnogram must not impose a great resistance in order to avoid pressure increase in the ML. A correct measurement can be obtained after a five-minute purge with high GF (10 L/min) in order to get rid of excessive moisture. As we can see the CO₂ removal is directly proportionate to GF (as in a NL it is proportionate to alveolar ventilation) and to P CO₂ in the blood entering the ML.
Sections of the ML which are ventilated, but not perfused, as it can happen in capillaries which are completely clogged with thrombosis, do not participate in the CO₂ removal. As with the NL, these sections are called dead space and it ratio can be calculated as follows:
Dead Space % = 100 (pCO₂, blood – PCO₂, gas) / pCO₂, blood.
In practice a failing CO₂ removal capacity causes an increase of GF in order to maintain a normal PaCO₂ in the patient.

Fig.2: An example of daily monitoring of Shunt% and Dead Space % of an oxygenator used for 40 days continuously.

Measuring pressures across the VV-ECMO circuit

As we can measure pressures across the pulmonary circulation by means of a Swan-Ganz Catheter we are able to measure in VV-ECMO pressures between the patient and the blood pump (drainage pressure) between the pump and the ML (oxygenator inlet pressure ) and beyond the oxygenator (oxygenator outlet pressure).

Fig.3 A brief scheme of the pressure monitoring of the ML and NL. SG, Swan Ganz Catheter. CVP, Central Venous Pressure. RV, Right Ventricle. PAP, Pulmonary Artery Pressure. LV,Left Ventricle. AP, Arterial Pressure.

The main interest in recording these pressure is obtaining the drainage pressure and the resistances across the ML. Drainage pressure is a negative pressure and, due to the risk of causing air emboli, it must be managed with great care and attention. It nevertheless gives us a fundamental information that can help us to prevent fluttering, with its abrupt fall in BF and ML oxygenating performance, cavitation, with the endothelial damage it can cause, and haemolysis.

The resistances of the ML can also be monitored (ML resistances = inlet pressure-outlet pressure/BF)and can give us paramount data on the functional status of the ML as resistances rise with thrombosis and capillary occlusion. ML resistance (or the pressure drop across the ML) is a very important adjunct to the oxygenating and decarboxylation parameters of the ML and can guide a correct timing of ML change.

Fig.4 An example of on line monitoring of pressure across the ECMO circuit

Recording data from the VV-ECMO circuit

The pressure data together with blood gas analysis from the ECMO inlet and ECMO outlet can give us a thorough perception of the performance of the ML. If a Swan-Ganz catheter is in place we can obtain mixed venous blood analysis and arterial gas analysis in order to evaluate the NL. Mechanical data from the NL such us respiratory compliance and resistance also are of great value.

In order to appreciate the complexity of these data most centres have an usually home-made electronic data sheet that collects raw data and calculated derived parameters. Obviously this kind of monitoring is not continuous and cannot give real time information though it is extremely important. Collecting these data can be cumbersome and is time consuming. It is nevertheless fundamental in giving a clear trend perception of the ML performance and the timing of a ML change and in giving information about the NL situation and its recovery eventually leading to the end VV-ECMO support.

Fig. 5 Example of the data electronic sheet used in our centre.

Many centres are experimenting ways of having a real-time continuous monitoring of the ML. The system our centre uses (Landing Real® Time Monitor, Eurosets S.r.L., Medola, Italy) is made of two units connected to the arterial and venous cannula, with spectrophotometric oximeters and infrared sensors which give temperature, Hgb value and SO2 both at inlet and outlet of the ML, and an ultrasonic flowmeter to measure BF. These values are connected to a monitor which also receives the pressure values of the circuit. The monitor doesn’t only show this values and their trends but also calculates and shows the derived data such as VO2 (consumed oxygen) DO2 (delivered oxygen) and the O2ER (oxygen extraction rate).

Fig.6 The two oximetric probes (on the left) and the flowmeter of our real time extracorporeal circuit monitoring device (Landing® Real Time Monitor, with permission of Eurosets S.r.L., Medolla, Italy).

Fig.7 Measured and calculated ECMO parameters obtained continuously (with permission of Eurosets S.r.L., Medolla, Italy).

These parameters are obtained without blood sampling and without accessing the extracorporeal circuit. Measured parameters must be calibrated daily when we perform our the ML inlet and ML outlet blood sampling and our thorough examination and analysis of the VV ECMO circuit.

The measured values are reliable and reasonably accurate but obviously the system has some limits. First of all it’s an adjunct to our daily VV-ECMO analysis but cannot take the place of it. Blood samples must be obtained daily also for calibrating the machine and we must refer to blood sampling every time we consider that the values have changed or the clinical situation does not mirror what we see on the screen. The system aims in casting some light in between our daily or one-per-shift analysis in order to avoid dramatic and unforeseen changes which could have been dealt with in more timely manner.

Another limit is that the system at the moment just focuses on the oxygenating function of our ML but not on CO2 removal. As these two functions sometimes do not fail at the same time this can be a major issue that the clinician must bear in mind. We expect a volumetric capnograph may be connected to the gas outlet of the ML and directly connected to the monitor; this way also V’CO2 may be calculated and monitored continuously.

We also must underline the importance of laboratory monitoring such as D-Dimer, CRP (C reactive Protein), aptoglobin, free blood haemoglobin and coagulation in getting a perspective of the inflammatory, thrombotic and haemolytic phenomena triggered by the extra-corporeal circuit. D-Dimer especially has been proposed as a good parameter of the ML aging and cannot be overlooked.

Fig. 8 : an example of daily monitoring of inflammatory and coagulation index of a single oxygenator used for 31 days continuously.

In conclusion real time monitoring of the VV ECMO circuit is an important adjunct and may enhance the safety of the technique in prompting more timely and accurate analysis of ML performance and avoiding unforeseen abrupt VV-ECMO failure. In future it may become the gold standard when also the CO2 removal will be monitored. Obviously also the function of the NL must be accurately analysed in order to remove the ML at the right time avoiding complications both from ECMO and ventilator.