Search

Two diffusion models have been developed to analyze gas transfer data previously measured in an intravascular artificial lung consisting of a central gas supply catheter from which are tethered a large number of blind-ended microporous fibers of equal length. A convective-diffusion model (CD) describes the countercurrent transfer of a binary gas pair when gas is supplied at constant pressure conditions, and a well mixed (WM) cycled pressure model predicts transfer when the gas supply pressure is time cycled between compression and vacuum conditions. Regression of gas to gas and liquid to gas excretion data with the CD model resulted in estimates of the liquid phase mass transfer coefficient kAI. Because these values were intermediate between the kAI expected for flow parallel to a cylinder and for flow normal to a cylinder, gas transfer was influenced by both the tethered region of the fiber that was nearly perpendicular to the axis of the test section and the free end of the fiber that rested along the wall of the test section. With a time cycled gas supply pressure, the enhanced carbon dioxide and oxygen excretion predicted by the WM model was similar to the data, but a loss in transfer efficiency with fiber length was not accounted for by the theory.

The woven tubes membrane oxygenator is a suitable configuration for the intracorporeal membrane oxygenator because of a high gas exchange performance and a compact packing of tubing. In this study the oxygen transfer performance of woven tubes was evaluated by an in vitro experiment with an external perfusion mode; the blood flow is outside of the tubes in order to reveal the feasibility of designing the intravascular oxygenator (IVOX) by the woven tubes. The oxygen transfer efficiency of the external perfusion mode is superior to that with the internal perfusion mode because of the larger convective mixing effect on the external surface of the tubes. Thus the use of the external perfusion mode results in the shorter necessary tube length for the rated condition, which enables making the oxygenator unit more compact. All of these features encourage the adoption of the woven tubes for use in the intravascular oxygenator.

Blind-ended, hollow fibers mounted on a pulmonary artery catheter may allow O2 and CO2 transfer in the vena cava, right ventricle, and pulmonary artery. The effects of fiber length, manifold number, and gas oscillation on mass and momentum transfer with water perfusate using mass spectrometry and mass flow controllers were studied. Manifolds with 112-196 microporous polypropylene fibers were mounted on 8 Fr multiple lumen, commercially available pulmonary artery catheters. Fiber lengths varied from 0.5 to 16 cm and surface areas from 7 to 220 cm2. Prototypes with 2 cm long fibers were constructed with 1-15 manifolds. A two manifold prototype with 8 cm long fibers and a surface area of 378 cm2 was also studied. The transfer failed to scale with manifold number because the steady gas flow was maldistributed to the manifolds. Oscillating gas pressures from 780 to 76 mmHg absolute at a rate of 40 cycles/min increased CO2 transfer up to 15-fold and O2 transfer up to 2.5-fold. Oscillation also corrected the maldistribution. Optimal fiber lengths of 3 and 1 cm for O2 and CO2, respectively, were seen with steady gas flow, and 8 cm for both with oscillatory gas flow.

It may be possible to design an intravascular membrane lung with gas transfer properties augmented by the natural flow oscillations in the venous and pulmonary circulation caused by the beating heart and ventilatory movements. The authors used a simple dye visualization technique, the Pierce-Donachy assist pump, and mass spectrometry to investigate these effects on membrane lungs made with tethered, blind-ended, microporous, polypropylene fibers using in vitro tests in water saturated with O2, CO2, and He. Prototypes were constructed on a 7.5 Fr pulmonary artery catheter. The fibers had an outer diameter (OD) of 380 microns and a wall thickness of 50 microns and were mounted on 4.8 mm OD sleeves. Control measurements were taken over a range of steady water flows from 0.4 l/min to 3 l/min. While pumping the same water flow rates with a roller pump, the Pierce-Donachy pump generated pulsatile flow at a rate of 45 beats/min and a systolic duration of 300 msec. This produced a phasic flow with an instantaneous average flow velocity varying from 0 to as high as 46 cm/sec. O2 and CO2 transfer increased by as much as 91% and 59%, respectively. The largest effects were seen at the lower water flow rates.

UC-PK1 cells were grown to confluence on microporous microcellulose membranes in order to test the feasibility of using transmembrane pressure (TMP) for controlling connective water transport in a bioartificial kidney. TMP was applied on either the apical or basal aspect of the polarized cell layer, while the fluid compartments on both sides of the membrane were perfused with tissue culture medium in a miniature flow chamber. The cell monolayer did not allow filtration in the apical to basal direction when positive TMP up to 30 mmHg was applied on the apical side. Application of positive TMP on the basal side led to measurable ultrafiltration. The hydraulic permeability, L(p), of the cell-seeded membranes was found to be increasing with time, reaching a steady state value after 60 min. There was a strong positive correlation between L(p) and applied TMP. For a constant TMP, L(p) was found to be independent of the shear rate between 2.6 and 10.5 sec-1. When positive pressure was changed abruptly back from the basal to apical site, the ultrafiltration rate decreased to zero within seconds. Morphologic studies suggest that the tight junctions between cells were broken by TMP applied on the basolateral aspect of the cell monolayer. It is concluded that, basolateral TMP may be used to control water flux in a bioartificial kidney.

An artificial pancreas consisting of beta cells cultured on synthetic semipermeable hollow fibers was tested in rats with alloxan-induced diabetes. When implanted ex vivo as arteriovenous shunts in the circulatory system these devices lowered concentrations of plasma glucose from 533 to between 110 and 130 milligrams per 100 milliliters, increased concentrations of plasma insulin, and restored intravenous glucose tolerance tests essentially to normal.

The combination of an ultrafiltration device with an exchanger whose semipermeable hollow fibres are covered with renal epithelial cells is proposed as a design for a bioartificial kidney. We first demonstrated that continuous ultrafiltration can be maintained for relatively long periods in the absence of anticoagulation. As a second step, we report here the feasibility of attaching and growing two lines of kidney epithelial cells (MDCK and LLC-PK1) on two different semipermeable materials, an acrylic copolymer and a polysulphone. Cells seeded on acrylic copolymer hollow fibres reach confluence within three weeks. Depending on the chemical and/or physical properties of the polymer, the cells show distinct differentiated morphology, which may influence their ability to perform specialized tasks.