Artificial Heart Valves

Artificial heart valves are either used as a replacement for human heart valves (prosthetic) or in cardiac assist systems (mechanical). The Biofluid Mechanics Lab is working mainly on heart valves for cardiac assist systems.

Clinical applications of cardiac assist systems continue to have a severe problem, namely thromboembolic complications. The problem originates mainly at the valves, which are usually made of an antithrombogenic material, such as bovine pericardium. However, the valve housing is made of a less suitable material, and wherever the blood flow is stagnant a thrombus is likely to form. Such stagnant blood flow is found in the space between the housing of the valve and the leaflets, called the sinuses. Consequently, thrombi often are generated in the sinuses.

The Biofluid Mechanics Lab is designing two new patented types of heart valves:

• S-shape valve
• Purge flow valve

In addition, the Biofluid Mechanics Lab is working on test methods for heart valves and a new method to measure the residence time of blood in heart valves called the fluorescent dye washout method.

S-shape valve

The S-shape valve (German patent no.: 196 04 881; EU patent no.: 971 02 039; US patent no.: 5 980 568) consists of a monoleaflet valve (grey) in a special designed duct (red). The duct is optimized according to optimum flow, which means that there is neither large flow acceleration nor stagnant areas.

S-shape valve

The flow in this valve was calculated by Computational Fluid Dynamics (CFD). You can find the results here. In addition, the flow was measured using a ten times enlarged model with Digital Particle Image Velocimetry (DPIV). These results can be found here.

 Top of the page

Purge flow valve

Principle: The purge flow valve is intended for use in cardiac assist systems. The special design (German patent no.: 198 07 599) reduces the formation of the stagnation zone behind the leaflets by means of a purge flow during systole. This purge flow is separated from the valve's main flow through a flow divider, thus directing a part of the main flow into the sinuses behind the leaflets.	Principle of the purge flow valve
Parameter study and design: The investigation and optimization of the purge flow effect was performed on a mono-leaflet valve due to the simple geometry. The devider's position and the geometry of the sinus and the devider were varied systematically. Details of the varied parameters are shown in the figure on the right. Theoretically, combining all possible parameter variations woul result in 200 models. Using a special factorial design technique known from quality management (Taguchi's method) the number of required models could be reduced to about 30. These models were designed using the 3D CAD Tool SolidWorks.	Varied parameters
Numerical investigation: To narrow down the choice, stagnation areas in the sinuses were computed using methods from CFD. The models with the smallest integrated stagnation areas were preselected and manufactured on a scale of 1:1. The figure shows a comparison of two different parameters — length of the leaflet and position of the flow divider — based on the calculated wall shear stresses. Areas of stresses lower than 0.5 Pa are marked blacked thus indicating unwanted separation and stagnation areas.	Numerical calculation (CFD)
Experimental investigation: The main hydrodynamic parameters were measured with a computer controlled valve tester and the washout of a dye previously filled into the sinus was observed, digitally recorded and quantified. In the figures on the right, a wash-out sequence and the course of the normalized gray value are shown. Subsequently the same valve geometries were investigated in an enlarged model — scale 2:1 — with DPIV in order to verify the CFD results. Both the numerical and the experimental investigation show that the best results are achieved with a short leaflet, a small sinus, a big flow divider and with the flow divider in symmetrical position. However, only one of the models with flow divider showed the expected large improvement of the washout process compared to the model without the flow divider (see the figure showing the course of the gray value). The DPIV investigation confirmed the results of the CFD and showed complex flow patterns in the sinus region. After further investigation the purge flow principle will be applied to the tri-leaflet valve.	Wash-out sequence Course of normalized gray value

 Top of the page

Testing method for artificial heart valves — bulk qualities

Numerous devices and mock-circulations have been described for the testing of artificial heart valves in regard to their pressure loss, closure time, closing and leakage volumes as well as energy losses. However, all devices have been troubled by the difficulty to generate and assess the precise flow through the valve and by the problem to define the arterial load, i.e. the artificial aorta.

The new test device (see figure) follows a radically different approach: there is no artificial ventricle with two valves, one of them being the test valve, instead, only a piston which forces the fluid through the test valve. Thus the movement of the piston defines the flow with great precision (0.3 %) and there is no influence from a second valve. As a result, there is no additional device needed to measure the flow. The piston is computer controlled and follows a physiological flow curve which is identical for all types of heart valves of the same size.

Valve tester

After the forward flow phase the controller switches over from flow control to pressure control, the piston moves slightly backwards, imitating the diastolic pressure difference between ventricle and aorta. This physiological pressure difference curve is mathematically defined and generated by the computer as well. Consequently there is no influence through an imprecisely defined after-load caused by a mechanically simulated elasticity of the aorta or peripheral resistance. Additionally the valve duct discharges into an open vessel. Since th transparent rigid aortic root is screwed in, this greatly simplifies the insertion and exchange of the test valve. This makes the tester suitable for production control.

The pressure difference across the valve is measured conventionally with two pressure transducers. The amplification of one transducer is changed according to the signal strength in order to achieve a higher resolution of the pressure signal during the systolic phase. The measurement of the piston displacement is done with a digital angular transducer.

The results — flow, pressure difference and energy loss — are printed out as curves, optionally as data lists. The output diagram includes the integrated data: closure time, closing volume, leakage volume, mean systolic pressure difference, closure time, energy loss during systole and diastole. It appears on the computer screen some seconds after the test.

Output diagram

The small size allows for the setup on a normal desk and as a result the preparation for a standard valve test is a question of some minutes.

Besides standard valve tests according to the ISO proposals, the device has also been used for the development of new valve designs and the investigation of failed explanted valves.

 Top of the page

Fluorescent dye washout method

For a long time, the hydraulic performance of artificial heart valves was the focus of interest. However, this does not really reflect the performance of the valve in the patient. Much more important are the thrombogenic qualities of the valve. How can these be assessed? While passing the valve, platelets are activated in areas of high shear stress and are likely to coagulate in stagnant areas. Flow separations behind the valve are an example of such areas of stagnant flow and the platelets may recirculate there many times. Given the critical shear rate and enough time, the coagulation starts and a thrombus is formed. The amount of time a platelet remains in the stagnant zone — known as residence time — is used as a measure of the thrombogenicity of the valve. The platelet stagnation zones of different valves have been investigated and their residence time calculated. In order to study the detailed flow behind the valve a 10:1 times enlarged model of the aortic valve was used. Dye was used as a model for the platelets. The flow was created in a pulsatile flow channel, the fluid used was water. To maintain Reynolds similarity the time scale was set to 1:254. In order to simulate a pulse rate of 70, one heart cycle lasted for 217 seconds. As a result, a common video camera was sufficient to study the flow. So that a realistic flow could be obtained a transparent model of the aortic root was placed in the flow channel. The area of interest was illuminated by a slide projector holding a slide with a narrow light slit. No intensive light sources were necessary, since the velocity was low and the video camera was very light sensitive. A fluorescent dye was mixed into the aorta. The higher the concentration of the dye the more light was reflected. The time dependant distribution of the concentration of the dye during the cycle was a measure of the platelets residence time behind the valve. The recorded cycle was then digitized to a PC and calculations were done using image processing software such as "Image" and custom made programs. The resolution of the pictures was 768 × 512 Pixel and used 256 grey levels. The following valves were investigated: Björk-Shiley Standard, Björk-Shiley Monostrut, Starr-Edwards Ball valve, St. Jude Medical, trileaflet PU valve, Jellyfish valve and a custom made ball valve.

We obtained video frames that showed areas of stagnant flow behind all the valves. In some valves the fluid was washed out better. There was a good correlation between the observed areas of flow stagnation and the thrombus formations found during post mortems. This proves, that the 10:1 flow channel using a fluorescent dye is a good method to use in order to obtain the residence time of different valve types and from that a measure of the thrombogenicity. One major advantage of the method is, that the individual cycle can be viewed and investigated and average velocity does not have to be calculated. In the future, evaluations will be done to decrease the thrombogenicity of artificial valves.

Two examples acquired with the described method are shown in the figures below.

Residence time of the fluid behind a bileaflet valve

Residence time of the fluid behind a tilting disc valve

 Top of the page