Frank R. Walters

Development of a low flow resistance intravenous oxygenator.

A potentially attractive support device for patients with acute respiratory failure is an intravenous membrane oxygenator. One problem, however, is that the membrane surface area required for sufficient gas exchange can unduly increase vena cavai pressure drop and impede venous return. The purpose of this study was to design and develop an intravenous oxygenator that would offer minimal venous flow resistance in situ. The device uses a constrained fiber bundle of smaller cross sectional size than the vena cava so as to effect an intentional shunt flow of venous blood around the fiber bundle and reduce the venous pressure drop caused by the device. A pulsating balloon within the fiber bundle redirects part of this shunt flow into reciprocating flow in and out of the fiber bundle. This offers dual advantages: 1) Blood flow through the fiber bundle is mainly perpendicular to the fibers; and 2) the requisite energy for driving flow comes largely

Recent progress in engineering the Pittsburgh intravenous membrane oxygenator.

The University of Pittsburgh intravenous membrane oxygenator (IMO) is undergoing additional engineering development and characterization. The focus of these efforts is an IMO device that can supply as much as one-half basal O2 consumption and CO2 elimination rates while residing within the inferior and superior vena cavae after peripheral venous insertion. The current IMO design consists of a bundle of hollow fiber membranes potted to manifolds at each end, with an intra-aortic type balloon integrally situated within the fiber bundle. Pulsation of the balloon using helium gas and a balloon pump console promotes fluid and fiber motion and enhances gas exchange. During the past year, more than 15 IMO prototypes have been fabricated and extensively bench tested to characterize O2 gas exchange capacity, balloon inflation/deflation over relevant frequency ranges, and the pneumatics of the sweep gas pathway through the device. The testing has led to several engineering changes, including redesign of the helium and sweep gas pathways within the IMO device. As a result, the maximum rate of balloon pulsation has increased substantially above the previous 70 bpm to 160 bpm, and the vacuum pressure required for sufficient sweep gas flow has been reduced. The recent IMO prototypes have demonstrated an O2 exchange capacity of as much as 90 ml/min/m2 in water, which appears within 70% of our design goal when extrapolated to scaled up devices in blood.

Respiratory dialysis. A new concept in pulmonary support.

Use of a new intravenous oxygenator made of hollow fiber membranes arranged around a centrally positioned balloon is reported. In vitro studies using fluorescent image tracking velocimetry and gas exchange analysis demonstrated enhanced convective mixing with balloon pulsations and augmented gas flux (100% increase in pO2) compared with the device in its static configuration. In vivo observations confirmed a greater than 50% increase in O2 flux with balloon activation. Those parameters that produce radial flow and convective mixing in vitro enhance gas flux in vivo, thus confirming the efforts to exceed the fluid limit translate into improved gas exchange.

Acute in vivo studies of the Pittsburgh intravenous membrane oxygenator.

The efficacy of an innovative intravenous membrane oxygenator (IMO) was tested acutely (6-8 hrs) in seven calves. The IMO prototypes consisted of a central polyurethane balloon within a bundle of hollow fibers with a membrane surface area of 0.14 m2. The IMO devices were inserted through the external jugular vein into the inferior vena cava of anesthetized calves (68.9 +/- 2.3 kg). Rhythmic balloon pulsation (60-120 bpm) was controlled with an intra-aortic balloon pump console. Oxygen sweep gas was delivered through the device at 3.0 L/min. Gas concentrations were monitored continuously by mass spectroscopy. The principal results were as follows: 1) oxygen and carbon dioxide exchange ranged from 125 to 150 ml/min/m2 and 150 to 200 ml/min/m2, respectively; 2) there was at least a 30-50% augmentation of gas exchange with balloon pulsation; 3) maximum exchange occurred with 60-90 bpm balloon pulsations; and 4) hemodynamic parameters remained unchanged. There were no device related complications, and the feasibility of insertion of the device by a cervical cut-down was established. These acute in vivo experiments show that the Pittsburgh IMO device can exchange oxygen and carbon dioxide gases in vivo at levels consistent with this current prototype design, and that intravenous balloon pulsation significantly enhances gas exchange without causing any end-organ damage.

Nitric oxide prevents human platelet adhesion to fiber membranes in whole blood.

During cardiopulmonary bypass or long-term extracorporeal life support, foreign surface induced platelet deposition in the oxygenator causes deterioration of gas exchange. In this study, the authors evaluated the effectiveness of nitric oxide (NO) in reducing the adhesion of platelets in whole blood to the surface of hollow fiber membranes. For this purpose, a test chamber was designed consisting of a gas exchanger with ten mitsubishi multi-layered composite hollow fibers (MHF: 257 mm OD; 203 mm ID; 70 mm length) and a polypropylene tube (16 mm OD; 100 mm length). Pure N2 (control) or nitric oxide (NO) (100 ppm, 200 ppm in N2) were delivered into the test chamber previously filled with 13 ml human whole blood. Platelet counts and platelet factor 4 (PF4), as a measure of platelet activation, were measured before and after either 1 or 2 hr of testing, and fibers were observed under scanning electron microscopic study (SEM) after each experiment. In the control and 100 ppm NO groups, platelet counts decreased and the level of PF4 increased during the 1 hr period. In the 200 ppm NO group, almost no platelet deposition could be observed on the surface of fibers under SEM. In conclusion, NO flow through hollow fiber membranes can markedly reduce platelet adhesion. Additional quantitative studies should define the optimal concentration for this effect and determine if this finding could improve oxygenator function, especially under conditions of long-term support.

A novel method for measuring hollow fiber membrane permeability in a gas-liquid system.

Designing an effective intravenous membrane oxygenator requires selecting hollow fiber membranes (HFMs) that present minimal resistance to gas exchange over extended periods of time. Microporous fiber membranes, as used in extracorporeal oxygenators, offer a minimal exchange resistance, but one that diminishes with time because of fiber wetting and subsequent serum leakage. Potentially attractive alternatives are composite HFMs, which inhibit fiber wetting and serum leakage by incorporating a true membrane layer within their porous walls. To evaluate composite and other HFMs, the authors developed a simple apparatus and method for measuring HFM permeability in a gas-liquid system under conditions relevant to intravenous oxygenation. The system requires only a small volume of liquid that is mixed with a pitched blade impeller driven by a direct current motor at controlled rates. Mass flux is measured from the gas flow exiting the fibers, eliminating the necessity of measuring any liquid side conditions. The authors measured the CO2 exchange permeabilities of Mitsubishi MHF 200L composite HFMs, KPF 280E microporous HFMs, and KPF 190 microporous HFMs. The membrane permeabilities to CO2 were 9.3 x 10(-5) ml/cm2/sec/cmHg for the MHF 200L fiber, 4.7 x 10(-4) ml/cm2/sec/cmHg for the KPF 280E fiber, and 2.8 x 10(-4) ml/cm2/sec/cmHg for the KPF 190 fiber. From these results it is concluded that 1) because of liquid-fiber surface interactions, the permeabilities of the microporous fibers are several orders of magnitude less than would be measured for completely gas filled pores, emphasizing the importance of measuring microporous fiber permeability in a gas-liquid system; and 2) the liquid diffusional boundary layer adjacent to the fibers generated by the pitched blade impeller is unique to each fiber, resulting in different boundary layer characterizations.