METHOD AND SYSTEM FOR MONITORING CARBON MONOXIDE (CO) ADMINISTRATION TO EX-VIVO FLUIDS

Abstract

The present invention relates to a method for combined administration of carbon monoxide (CO) to an ex-vivo fluid and monitoring of the carbon monoxide administration, said method comprising: (i) generating CO by reacting a CO releasing molecule (CORM) with a release triggering molecule; (ii) administering CO to an ex-vivo fluid by contacting the ex-vivo fluid with the CO generated in step (i) via a gas-permeable membrane; (iii) analyzing carbon monoxide and/or a carbon monoxide marker after administering in step (ii) CO to the ex-vivo fluid by complementary monitoring techniques; (iv) adjusting the CO administration based on the analysis of the carbon monoxide or the carbon monoxide marker carried out in step (iii), if necessary. It furthermore relates to an extracorporeal circuit system for use in the method of the invention.

Claims

1. A method for the combined (a) administration of carbon monoxide (CO) to an ex-vivo fluid and (b) monitoring of said carbon monoxide administration, said method comprising the steps of: (i) generating CO by reacting a CO releasing molecule (CORM) with a release triggering molecule, (ii) administering CO to said ex-vivo fluid by contacting the ex-vivo fluid with the CO generated in step (i) via a gas-permeable membrane, (iii) after administering CO to the ex-vivo fluid in step (ii) analyzing carbon monoxide and/or a carbon monoxide in said ex-vivo fluid by complementary monitoring techniques, (iv) based on the analysis of the carbon monoxide or the carbon monoxide marker carried out in step (iii), optionally adjusting the CO administration by repeating step (ii) if necessary.

2. The method according to claim 1, wherein the ex-vivo fluid is either blood obtained from a subject or a perfusion liquid.

3. The method according to claim 1, wherein the carbon monoxide marker is carboxyhemoglobin (COHb).

4. The method according to claim 1, wherein the complementary monitoring techniques utilized in step (iii) include at least two techniques selected from the group consisting of: (i) COHb measurement in a blood sample of a subject, (ii) measuring the CO concentration in the exhaled breath of a subject, and (iii) measuring the CO concentration in the exhaust air of an oxygenator, wherein the oxygenator is integrated in a cardiopulmonary support system configured to be connected with the blood cycle of a subject.

5. The method according to claim 1, wherein the complementary monitoring techniques applied in step (iii) are supported by machine learning-based CO level prediction.

6. An extracorporeal circuit system for use in administering CO to an ex-vivo fluid by the method according to claim 1, said extracorporeal circuit system comprising: an extracorporeal carbon monoxide releasing system (ECCORS) that includes a membrane module having an outer compartment and an inner compartment, wherein: (i) the outer compartment is connected to a primary circuit carrying blood of a subject, (ii) the inner compartment is connected to a secondary circuit in which CO is generated by reacting a CO releasing molecule with a release triggering molecule; and (iii) the inner and outer compartments in the membrane module are separated from each other by a gas-permeable membrane that allows CO permeation generated in the secondary circuit from the inner compartment into the outer compartment, thereby administering CO to the ex-vivo fluid, further wherein said primary circuit is configured to be connected with the blood cycle of a subject.

7. The extracorporeal circuit system according to claim 6, wherein the inner compartment and the outer compartment of the membrane module are separated by tube membranes.

8. The extracorporeal circuit system according to claim 6, wherein the primary circuit further comprises an oxygenator and a pump.

9. The method according to claim 1, wherein the gas-permeable membrane is a silicone membrane or a PTFE (polytetrafluoroethylene) membrane.

10. The method according to claim 1, wherein the CORM is a metal carbonyl compound.

11. The method according to claim 1, wherein the CORM is a metal carbonyl compound selected from the group consisting of Mo(CO).sub.3(CNCH.sub.2COOH).sub.3 (Beck1), Mo(CO).sub.3(CNCH.sub.2CONaO).sub.3 (Beck1-Na), CORM-ALF794, CORM-1, CORM-2, CORM-3, and CORM-401.

12. The method according to claim 1, wherein the release triggering molecule is selected from the group consisting of a sulfur containing compound, a nitrogen containing compound, and an oxidizing compound.

13. The method according to claim 1, wherein the release triggering molecule is an oxidizing compound selected from the group consisting of iron(III)chloride (FeCl.sub.3), potassium permanganate (KMnO.sub.4), cer(IV)sulfate (Ce(SO.sub.4).sub.2), potassium dichromate K.sub.2Cr.sub.2O.sub.7, gold(III)chloride (AuCl.sub.3), and silver nitrate (AgNO.sub.3).

14. The method according to claim 1, wherein the CORM is added to an aqueous solution of the release triggering molecule or the release triggering molecule is added to an aqueous solution of the CORM.

15. The extracorporeal circuit system according to claim 6, wherein the extracorporeal circuit system is part of a cardiopulmonary support system or a dialysis system.

16. The extracorporeal circuit system according to claim 6, wherein the gas-permeable membrane is a silicone membrane or a PTFE (polytetrafluoroethylene) membrane.

17. The extracorporeal circuit system according to claim 6, wherein the CORM is a metal carbonyl compound.

18. The extracorporeal circuit system according to claim 6, wherein the CORM is a metal carbonyl compound selected from the group consisting of Mo(CO).sub.3(CNCH.sub.2COOH).sub.3 (Beck1), Mo(CO).sub.3(CNCH.sub.2CONaO).sub.3 (Beck1-Na), CORM-ALF794, CORM-1, CORM-2, CORM-3, and CORM-401.

19. The extracorporeal circuit system according to claim 6, wherein the release triggering molecule is selected from the group consisting of a sulfur containing compound, a nitrogen containing compound, and an oxidizing compound.

20. The extracorporeal circuit system according to claim 6, wherein the release triggering molecule is an oxidizing compound selected from the group consisting of iron(III)chloride (FeCl.sub.3), potassium permanganate (KMnO.sub.4), cer(IV)sulfate (Ce(SO.sub.4).sub.2), potassium dichromate K.sub.2Cr.sub.2O.sub.7, gold(III)chloride (AuCl.sub.3), and silver nitrate (AgNO.sub.3).

Description

DESCRIPTION OF THE FIGURES

[0094] FIG. 1A shows a schematic drawing of the membrane module being part of the extracorporeal CO releasing system (ECCORS).

[0095] FIG. 1B shows a schematic application of the extracorporeal CO releasing system (ECCORS) in an exemplary veno-arterial circuit of an extracorporeal life support system.

[0096] FIG. 2 shows the course of COHb (%) measured by blood gas analysis (BGA) and pulse oximetry (SpCO) as well as the concentration (ppm) of carbon monoxide exhaled from the lungs and discharged from the oxygenator. Average initiation of ECCORS treatment was at t=24.24.7 minutes after start of asphyxiation (t=0). N=7; meansSD.

[0097] FIG. 3 shows in (A) a scatter plot and correlation of COHb measurement from BGA to exhaled CO from the lungs showing a correlation coefficient of R.sup.2=0.43; in (B) a scatter plot and correlation of COHb measurement from BGA to exhausted CO from the oxygenator showing a correlation coefficient of R.sup.2=0.59; in (C) a scatter plot and correlation of COHb measurement from blood gas to non-invasive SpCO measurement using pulse oximetry with a coefficient of R.sup.2=0.37

[0098] FIG. 4 shows an example of an algorithm controlling extracorporeal carbon monoxide releasing system connected to an extracorporeal circuit system according to the invention. Systemic levels of CO are constantly assessed using four complementary monitoring techniques including COHb analysis from blood (not online), pulsoxymetric analysis of COHb, CO quantification in exhaled breath as well as exhaust of the oxygenator. These parameters are processed in an actual-target comparison approach (i) controlling the CO release rate and hence systemic CO level within the patient (including but not limited to proportional-integral-derivative controlled modulation of the kinetic between a release trigger and a CO-releasing molecule within the extracorporeal carbon monoxide releasing system) and (ii) inducing an emergency bypass of the device in case of unforeseen overdoses.

[0099] FIG. 5 shows a heat map of Pearson r correlation coefficients (absolute value) between influential parameters for COHb prediction from respiratory, acid-base and hemodynamic parameters (SpCO: COHb measured from pulse oximetry; FiO2: fraction of inspired oxygen; ECLS: extracorporeal life support; PCWP: pulmonary capillary wedge pressure; SVR: systemic vascular resistance).

[0100] FIG. 6. To improve analytical precision when analyzing systemic CO levels in real-time, continuous monitoring parameters shown in FIG. 3 were processed using Random Forest regression and analyzed in randomized 5-fold cross-validations. Predictions are plotted vs. experimental COHb values: A) Three continuous, real-time parameters, B) only CO quantity in the exhaust of the oxygenator [ppm]. Solid lines indicate perfect predictions, dashed lines show the 90% prediction interval.

EXAMPLE 1

[0101] The method and system as disclosed herein was designed for easy implementability in clinical routine. Medical-translational significance thereof is demonstrated following system development by implementing the system into a common extracorporeal circuit system setup used for preclinical testing in swine. A schematic application of a venoarterial extracorporeal circuit system comprising an extracorporeal carbon monoxide releasing system (ECCORS) for extracorporeal life support is shown in FIG. 1B.

[0102] After obtaining consent by Regierungspraesidium Freiburg, seven landrace-hybrid pigs (weight 522 kg) were anesthetized (continuous infusion of propofol, fentanyl and cisatracurium), mechanically ventilated and monitored with electrocardiogram, pulse-oximetry and invasive blood pressure measurement by cannulating the carotid and pulmonary artery. Instrumentation with veno-arterial ECMO (extracorporeal membrane oxygenation) was carried out in a sterile technique: Venous access was achieved by ultrasound-guided cannulation of femoral vein (21 French HLS cannula, Maquet, Rastatt, Germany), while arterial inflow into the femoral artery was established using 17 French HLS cannula from Maquet, (Rastatt, Germany). The ECMO circuit (primary circuit) was then set-up using a customary centrifugal pump head (Revolution 5, Sorin, Rome, Italy), an oxygenator (EOS ECMO, Sorin, Rome, Italy), and 3/8 tubing for interconnection. The system was regulated using a modified console (core element: SCP-Console, Sorin, Rome, Italy).

[0103] The ECCORS was then integrated in line with the oxygenator into the primary ECMO circuit. A commercially available PDMSXA-1.0 silicone membrane module from PermSelect (AnnArbor, II) served as core functional element of the ECCORS. Silicone tube membranes within the PDMSXA module (inner diameter 190 m; outer diameter 300 m) separate the outer blood-carrying compartment (primary ECMO circuit) and an inner CORM-carrying compartment (secondary circuit). The outer compartment was connected to the 3/8 tubing of the primary ECMO circuit (vide supra). The inner compartment was connected to the secondary circuit comprising a solution of 15 g FeCl.sub.3 in 0.04 L water. The tubing of the secondary circuit was equipped with a port for CORM injection as well as two gastight bags (Reservoir bag 1 L, Sorin, Rome, Italy) for preventing overpressure in the system. The secondary circuit, was constantly circulated in tubing using a Masterflex Console Drive from Cole-Parmer (Vernon Hills, II).

[0104] FIG. 1A shows a schematic drawing of a membrane module being part of the extracorporeal CO releasing system (ECCORS). To address safety challenges associated with pressurized CO cartridges, CO within ECCORS is generated on demand by injecting the CO releasing Molecule (CORM) Beck1 into the iron-CORM circuit, containing the trigger substance FeCl.sub.3. After CO generation only CO (but no other constituent) permeates the gas permeable silicon hollow fibers within ECCORS and is hence delivered to the systemic circulation. The surface of the silicone hollow fibers inside the module determine the COHb concentration in steady state.

[0105] CO release was initiated by an injection of a solution of 1 g Beck1 (Mo(CO).sub.3(CNCH.sub.2COOH).sub.3) in 0.02 L water into the secondary circuit. Further injection of Beck1 was controlled by feedback measurement to the desired target and hence based on monitoring parameters detailed in FIG. 2 and FIG. 3, vide infra.

[0106] Liberated CO gas circulated in the tubing and excess gas was contained in two impermeable gas bag reservoirs connected to the circuit, enabling diffusion to the closed-loop tubing. Consequently, CO was allowed to permeate from the CORM into the blood compartment, but permeation of the non-gaseous constituents was inhibited by the membrane.

[0107] CO was quantified in the oxygenator's exhaust air, and the exhaled air using two MX6 iBrid (Industrial Scientific, USA) measurement devices. COHb was quantified using blood gas analysis (cobas b 123, F. Hoffmann-La Roche, Switzerland) and multi-wavelength pulse oximetry (Rad-97, Massimo, USA).

[0108] After reaching the therapeutic CO level, steady state was maintained by controlled injections of Beck1 into ECCORS under tight feedback control of online monitoring parameters including CO in exhaled air, as well as CO in the exhaust of the oxygenator (FIG. 2). COHb analysis from blood samples was analyzed offline (see FIG. 2). This technique is the current gold standard for assessing systemic CO levels, but automatic online measurement (and hence implementation in an automatic feedback loop, vide infra) cannot be performed.

[0109] COHb formation was also assessed using a pulse oximeter (continuous technique), however, as this technique is primarily used by medical first responders to approximate COHb values, the data only moderately correlated with the COHb as analyzed from blood (R.sup.2=0.37, FIG. 3C).

[0110] The CO level in exhaled air, as well as the CO level in the exhaust of the oxygenator was also assessed and likewise only correlates moderately to COHb as analyzed from blood samples (R.sup.2=0.43, and 0.59, respectively, FIGS. 3A and 3B). Therefore, all three continuous CO-monitoring approaches alone are not appropriate to reliably distinguish between therapeutic levels (10% CO-Hb, and borderline toxicitye.g. 16% CO-Hb)which is one challenge to be solved by this invention.

[0111] The data showed that no toxic components of CORM, especially no transition metals can be found in the tissue or blood. Tissue was analyzed for CO effects and residual CORM constituents in the harvest organs and blood using ICP-MS analysis for molybdenum. A separation of two circuits is accomplished, such that CO is generated independently from blood flow and in a controlled fashion. By dividing two separate circuits with a solely CO-permeable, but otherwise impervious membrane, it is possible to protect organs from ischemia-reperfusion-injury by administering CO into an extracorporeal circuit without exposure of toxic CORM components to these organs.

Example 2

[0112] A machine learning based approach was established and validated in order to increase the analytical strength of the systemic CO monitoring protocol. This approach processes the real-time CO monitoring parameters and provides a predicted CO level value for subsequent modification of the feedback loop. This protocol allows improving the analytical power of the delivery approach thereby significantly increasing the translational significance of the concept.

[0113] Method:

[0114] Machine learning models were generated using Random Forest regression (Breiman L. Random Forests. Machine Learning. 2001; 45(1):5-32) as implemented in the python library scikit-learn (Pedregosa F, Ga, #235, Varoquaux I, Gramfort A, Michel V, et al. Scikit-learn: Machine Learning in Python. J Mach Learn Res. 2011; 12:2825-30). The number of estimators was set to 1000. The predictive power of the models was assessed in randomized 5-fold cross-validations and in leave-one-animal-out cross-validations.

[0115] Results:

[0116] During in-vivo studies, multiple parameters (see FIG. 5) were plotted and analyzed, accordingly. In correlation studies, the factors CO Lungs, CO Oxygenator, and SpCO (carboxyhemoglobin measurement from pulse oximetry) were identified as most influential for COHb prediction with R.sup.2 values of 0.43, 0.59, and 0.37, respectively (see FIGS. 3 A, B, C).

[0117] In an effort to increase accuracy and precision of the analytical setup, we used these parameters to generate a multifactorial COHb prediction model. As baseline, the predictive power of a multiple linear regression (MLR) approach was assessed in a randomized 5-fold cross-validation (CV) and a leave-one-animal-out CV, respectively. The three-parameter MLR model predicted COHb values with a mean absolute error (MAE) of 2.36% COHb and a Q.sup.2 (R.sup.2 of CV prediction results) of 0.45, suggesting moderate prediction quality (see Table 1). The leave-one-animal-out CV indicates poorer performance of the MLR model with an MAE of 2.82% and a Q.sup.2 of 0.10. To improve the quality of COHb predictions, non-linear machine learning models were built using Random Forest regression (RF). Using RF, the COHb prediction power increased to an MAE of 1.35% at a Q.sup.2 of 0.79 in the randomized 5-fold CV and an MAE of 1.41% at a Q.sup.2 of 0.72 in the leave-one-animal-out CV. The randomized 5-fold CV indicates a 90%-prediction interval of 2.80% COHb (FIG. 6A). Random forest regression was repeated with CO Oxygenator as the only input parameter resulting in a prediction model with an MAE of 2.16% at a Q.sup.2 of 0.52 in a randomized 5-fold CV (FIG. 6B) and an MAE of 1.99% at a Q.sup.2 of 0.46 in a leave-one-animal-out CV.