Carbon dioxide transfer in membrane oxygenators and associated membranes
A recently developed therapy for treatment of acute respiratory failure requires that the patient's metabolic carbon dioxide production be eliminated by a membrane oxygenator operated in an extracorporeal blood circuit. In conjunction with peripheral cannulation, the oxygenator should be optimised for CO₂ removal at low blood flow rates of 1.5 ℓ/min or less for adults. An extensive literature survey revealed that very few publications dealt with oxygenator CO₂ performance at low flow rates. Two commercial devices, the Terumo CAPIOX II (1.6 m² and 3.3 m² membrane areas) hollow fibre oxygenator and the Travenol TMO (2.25 m² membrane area) parallel-plate oxygenator were evaluated in relation to the new therapy. A theoretical model describing carbon dioxide transfer in membrane oxygenators was used to correlate the experimental data. The Terumo CAPIOX II 3.3 m² unit was the only device capable of satisfying the carbon dioxide removal requirements necessary for the new therapy at the low blood flow rates stipulated. Effects of blood and gas flow maldistribution were also studied in the TMO and CAPIOX II units respectively. Non-uniform blood flow was not a major factor contributing to the decline in CO₂ transfer performance compared with theory. This was confirmed in experiments with a modified TMO unit. Comparison with theory indicated that the membrane resistance was the controlling factor for CO₂ transfer in the CAPIOX II device. A method was developed to assess the CO₂ transmission rate (Gco₂) through oxygenator membranes under gas-membrane-liquid contact conditions. This forms the basis for the selection of suitable membrane materials for oxygenators. Although the GCO₂ values for homogeneous silicone rubber membranes were consistent with the results of previous workers, significantly higher values were obtained for microporous polypropylene membranes. For microporous membranes under liquid contact conditions a 5-fold reduction in GCO₂ is obtained in this study compared to gas-membrane-gas tests, indicating that micropore wetting imposes a significant resistance to CO₂ transfer.