`Helmut Reul, Klaus Potthast
`
`Helmholtz-Institute for Biomedical Engineering, Aachen, Germany
`
`Background and aims of the study: The current stan-
`dards for accelerated heart valve testing have consid-
`erable differences in test conditions. Another prob-
`lem arises from the fact that such test systems are not
`standardized at all. It was shown earlier that different
`
`test systems generate totally different valve loading,
`even if operating at standard conditions. The present
`study aimed to improve this unsatisfactory situation
`and to develop a new concept where actual loading of
`valves is measured either in vitro or in vivo under
`
`physiologic conditions and subsequently to repro-
`duce these conditions during accelerated testing.
`Methods: Integral loading forces at valve closure were
`measured for several valve types using a piezoelectric
`force ring within a real-time circulatory mock loop
`under physiologic conditions. This facilitated defini—
`tion of a physiologic loading range. Physiologic load-
`ing was subsequently reproduced in a single-chamber
`accelerated test system. Working conditions obtained
`in terms of stroke, bypass flow and compliance served
`as design criteria for a new test chamber and a com-
`
`plete 12-chamber accelerated testing system.
`Results: The integral loading obtained using the force
`ring showed a correlation with previous in vitro and
`in vivo results of strain-gauged valves. Loading
`forces for mechanical valves are about one order of
`
`magnitude higher than for bioprosthetic valves and
`are strongly related to cardiac output for both valve
`types. At physiologic loading, however, the differen-
`tial pressures across the valves are considerably
`below those given in FDA guidelines.
`Conclusions: This pilot study demonstrates that phys-
`iologic valve loading is reproducible over a wide
`range under appropriate testing conditions. It also
`showed that, at the back-pressures of the currentstan-
`dards, the loading forces during accelerated testing
`exceed the real-time loading forces by far and, thus,
`may provide unrealistically high valve loads. These
`initial findings indicate that amendments of the cur-
`rently valid standards may be need to be accorded.
`
`The Journal of Heart Valve Disease 1998;7t151-157
`
`In 1996 we published a critical review on the state of
`the art of heart valve wear and fatigue testing as repre-
`sented by the current standards of ISO, FDA and CEN
`(1). These standards have considerable differences in
`
`test conditions as far as pressure differences across the
`test valves and total cycle numbers are concerned.
`Moreover, since the testing devices themselves are not
`standardized, different devices generate totally differ-
`ent loading conditions on the valves, even if the tests
`are carried out under the same standard. This study
`clearly demonstrated that, under the current stan-
`~ dards, the conditions of actual, in vivo impact loading
`of a valve cannot be reproduced. Also, the two test
`devices which were compared, although both operat-
`ing at the given standards conditions, generate totally
`different loading conditions. For one tester loading
`decreases with increasing cycle rate, while for the other
`it increases. For all testers, time history of loading and
`
`Address for correspondence:
`Prof. Dr.-lng. H. Rcul, Helmholtzlnstitute for Biomedical Engineer-
`ing, Pauwelsstr. 20, D-52074 Aachen, Germany
`
`pressure difference across the valve are a function of
`test frequency, compliance and other factors.
`To improve this unsatisfactory situation it was
`strongly suggested that the actual loading conditions
`of each valve type be measured either in animal mod-
`els in vivo or within a circulatory mock loop which pro-
`vides physiologic loading conditions;
`this loading
`could then be subsequently reproduced during accel-
`erated wear testing. The present study is set out to
`investigate this concept.
`
`Measurements of valve loading
`under physiologic conditions
`
`Since most valves - and especially bioprostheses - are
`not suitable for the attachment of strain gauges due to
`the lack of appropriate fixation points which are repre-
`sentative for valve loading, a piezoelectric force mea—
`surement
`ring (Kistler,
`type 906 1A, Winterthur,
`Switzerland) (Fig. 1, top right) was used to assess load-
`ing at valve closure. This type of force transducer mea-
`sures the integral loading force acting on the valve.
`
`© Copyright by lCR Publishers 1998
`
`PAGE 1 OF 7
`
`WATERS TECHNOLOGIES CORPORATION
`
`EXHIBIT 1014
`
`WATERS TECHNOLOGIES CORPORATION
`EXHIBIT 1014
`
`PAGE 1 OF 7
`
`
`
`152 Durability/wear testing of heart valve substitutes
`H. Reul, K. Potthast
`
`J Heart Valve Dis
`Vol. 7. No. 2
`March 1998
`
`semiconductor
`pressure transducer
`
`( NF : 100 kHz )
`
`
`
`
`
`inductive
`displacement
`transducer
`
`elektrornagnetic
`vibrator
`W
`
`
`
`Figure 1: Schematic of single-chamber accelerated valve test
`system with piezoelectric force ring in the upper right.
`
`The next step was to correlate these integral loading
`forces with previously obtained in vivo and in vitro
`results (2,3) measured by a strain-gauged 29 mm Bjork—
`Shiley Convexo-Concave (BSCC) mechanical
`tilting
`disk valve. For this purpose, the strain—gauged BSCC
`valve was mounted on the force ring and inserted into
`the mitral position of a physiologic circulatory mock
`loop as previously described (4). Both, strut loads and
`integral loading force were measured simultaneously.
`For these measurements the left ventricular dp/dt
`(according to the FDA guidelines averaged over the
`last 20 ms before valve closure) of the model ventricle
`
`was varied between 500 and 1500 mmHg/s by increas-
`ing the stroke volume at a fixed test rate of 70 per min,
`a mean aortic pressure of 100 mmHg, and a systolic
`duration of 15%.
`
`The results are shown in Figure 2. The upper curve
`represents the force ring measurements which range
`from 25 to 110 N integral loading force at the corre-
`sponding dp/dt values. The two lower curves repre-
`sent the strut loading forces obtained by strain gauges
`in vivo and in vitro for the same valve; these range
`from 2 to 20 N strut loading force. The difference in
`measured forces between the two methods is obvious
`
`and is related to the fact that the strain gauges measure
`only loading of a single strut caused by strut deflection,
`while the force ring measures the integral total loading
`force acting on the closed valve. Nevertheless, both
`
`PAGE 2 OF 7
`
`120
`
`100
`
`z 80
`E2
`u? so
`
`40
`
`20
`
`Integral Loading Force in Vitro
`(Piezoelectric Force Ring)\
`
`Strut Loading in Vitro
`
`Strut Loading in Vivo
`
`0
`
`500
`
`2000
`1500
`1000
`Left Ventricular dp/dt/ mmHg/s —>
`
`2500
`
`Figure 2:111 vivo and in vitro loading of a 29 mm BSCC valve
`as a function of left ventricular dp/dt before valve closure.
`
`methods give a good correlation with left ventricular
`dp/dt and, thus, can as well be correlated among each
`other.
`
`Measurements under accelerated testing condi-
`tions
`
`‘
`
`After establishing the above correlation between inte-
`gral forces and left ventricular dp/dt the valve was
`inserted into a specifically developed single-chamber
`system for accelerated valve testing (also shown in Fig.
`1).
`
`An electromagnetic vibrator (Ling Dynamics, type
`409, Royston, UK) generates a sinusoidal flow through
`the test valve by compression and extension of a metal-
`lic bellows. When the valve is closed, fluid from the
`
`lower chamber flows to the upper chamber via an
`adjustable bypass. This bypass serves simultaneously
`to control the pressure difference across the valve,
`which is measured upstream and downstream of the
`valve by two semiconductor pressure transducers with
`a natural frequency of 100 kHz (Cobe Disposable
`Transducer, Lakewood, CO, USA). Test fluid is water
`at room temperature.
`The test rate was increased from 200 to 2000 per min
`and the integral loading forces were measured. The
`pressure difference across the valve was kept constant
`at 120 mmHg for all test rates. The results are present-
`ed in Figure 3.
`The shaded area represents the range of loading
`forces under all potential physiologic conditions as
`obtained in vivo and in vitro and as depicted in Figure
`2. The central curve shows the measured integral load-
`ing forces under accelerated testing conditions under
`observance of the FDA conditions for accelerated valve
`
`testing such as full opening and closing and 120 mmHg
`pressure difference. The minimally adjustable loads
`
`l
`r
`
`
`
`,
`
`PAGE 2 OF 7
`
`
`
`I Heart Valve Dis
`Vol. 7. No. 2
`March 1998
`
`140
`
`_. ND
`
`.n OO
`
`IntegralLoadingForce/N—> 88
`
`
`
`
`
`3.O
`
`ND
`
`Minimally Adjustable Loading
`
`Durability/wear testing of heart valve substitutes
`H. Reul, K. Potthast
`
`153
`
`
`
`2L/min
`
`7L/min
`
`Force/N——>
`
`
`
`2L/min
`
`7L/min
`
`Figure 4: Integral loading forces for a 27 mm SIM mechanical
`valve and a 25 mm ISP bioprosthesis measured under real—
`time conditions within a circulatory mock loop.
`
`This range of test conditions represents the normal
`physiologic range and can be used for the characteriza-
`tion of corresponding loading conditions. For each
`valve type the measured loading forces were averaged
`over 100 cycles.
`The results (Fig. 4) show that loading for the mechan—
`ical valve is about one order of magnitude higher than
`for the bioprosthetic valve, and loading increases with
`a factor of about two when cardiac output changes
`from 2 l/min to 71/min.
`
`Design of a new fatigue tester
`
`The principal design criteria for a valve test chamber in
`terms of stroke, displaced volume, bypass flow and
`compliance were obtained by means of the above-men-
`tioned single-chamber system (see Fig. 1). Based on
`these criteria a new test compartment for general use
`was designed (shown schematically in Fig. 5). For the
`new design, the electromagnetic vibrator was replaced
`by a swash plate with adjustable stroke. Otherwise, the
`operating principle was the same as already described
`above.
`
`Larger valves generate higher impact loading forces
`than smaller ones. Therefore, since the new test com-
`partment is intended for use with all kinds of valve
`sizes, an adjustable compliance chamber was added for
`additional control of
`loading forces. The desired
`impact force can then be adjusted by varying the fol-
`lowing four parameters:
`
`- stroke of the swash plate;
`—
`test rate;
`
`- air compliance; and
`- bypass throttle adjustment.
`
`The FDA guidelines require that for all valves, test-
`ing should be conducted on three of the largest, medi-
`um and smallest of each valve type. One equivalent tis-
`
`
`
`V
`
`o0
`
`200
`
`400
`
`600
`
`1000 1200 1400 1600 1800 2000 2200
`800
`Test rate / beats / min —>
`
`Figure 3: Integral loading forces within the single chamber
`accelerated test system obtained for a 29 mm BSCC valve at a
`hack—pressure of 120 mmHg (FDA—condition).
`
`are represented by the lower curve. Higher loads can
`be easily generated by adjusting stroke and bypass
`throttle of the test system. These results clearly show
`that physiologic valve loading can be reproduced over
`a wide range under appropriate accelerated testing
`conditions. It is also evident that, under FDA condi-
`tions, only a single loading force which increases with
`higher test rates can be generated for a defined test rate,
`whereas test conditions outside of the current standard
`
`in terms of pressure difference facilitate the adjustment
`of a range of physiologic loading conditions at any test
`rate. Thus, valve loading at resting conditions as well at
`exercise conditions can be adjusted, resulting in a much
`better simulation of the loading history of an implant-
`ed valve.
`
`Valve loading forces for alternative valve types
`
`So far, all presented results are only valid for one single
`valve type and size, a 29 mm diameter BSCC tilting
`disk valve. Therefore, in order to develop a generally
`valid concept, it is necessary to extend the above find-
`ings to other valve types and sizes. For this purpose, a
`St. Jude Medical (SIM) 27 mm mechanical bileaflet
`valve and an Ionescu—Shiley 25 mm pericardial bio—
`prostheses were selected as a first step. Both valves
`were inserted into mitral position of our circulatory
`mock loop and integral loading forces were measured
`under the following experimental conditions (accord-
`ing to FDA guidelines for pulsatile flow valve testing):
`
`Cardiac output: 2 l/ min and 7 l/min;
`Test rate: 70 per min;
`Mean aortic pressure: 100 mmHg;
`Mean atrial pressure: 10 mmHg; and
`Systolic duration: 35%.
`
`PAGE 3 OF 7
`
`PAGE 3 OF 7
`
`
`
`154 Durability/wear testing of heart valve substitutes
`H. Real, K. Potthast
`
`Test compartment of HlA-FT2
`
`7*" 7
`
`Observation Tube
`
`, / Test-Compartment
`
`.: '2» Pressure Taps
`
`, Compliance Chamber
`
`open to atmosphere '
`
`Throttte ,
`
`Testvalve
`Bypass
`
`.7
`
`Piston Rod
`
`
`
`
`)B r“ * Adjustablesn'oke
`Smash Plate
`
`Figure 5: Schematic cross-section of newly developed test com-
`partment for accelerated valve testing.
`
`sue annulus diameter of each type reference valve must
`be tested under identical conditions. This results in a
`
`total of 12 test valves. Accordingly, a 12-cylinder
`fatigue tester (HIA-FTZ) has been designed and manu—
`factured (Fig. 6).
`The first two of the above-listed parameters cannot
`be adjusted individually; they are the same for all 12
`test compartments. Thus, after the correct adjustment
`of one compartment the other 11 must be adjusted by
`variation of air compliance and bypass throttle flow.
`An example (Fig. 7) shows where the peak loading
`forces for a 27 mm SIM valve have been varied by
`changing the air compliance at a constant test rate,
`bypass throttle and stroke settings. Peak load decreas-
`es linearly with increasing compliance volume and can
`easily be adjusted for physiologic loading conditions.
`
`Verification of concept within new fatigue tester
`
`In a final step a 27 mm mechanical SIM valve and a 25
`mm Ionescu-Shiley pericardial valve (ISP) were inserted
`
`PAGE 4 OF 7
`
`I Heart Valve Dis
`Vol. 7. No. 2
`March 1998
`
`
`
`Figure 6: Top view of newly developed 12-chamber accelerated
`test system (HIA—FTZ).
`
`into the new test compartment and the loading forces
`together with the pressure difference across the valves
`were measured within the fatigue tester at a test rate of
`1000 per min for the SIM and 600 per min for the ISP,
`respectively. The results (Figs. 8 and 9) show that for
`both valve types the peak loading forces are within the
`previously determined physiologic range (compare Fig.
`4). The differential pressures, however, are considerably
`below the pressures given in the FDA guidelines, which
`should be adjusted to 120 mmHg for both valve types.
`An example illustrating this controversy for the SIM
`valve (Fig. 10) shows that, in this case, the pressure dif-
`ference across the valve was adjusted to 120 mmHg by
`reducing the test chamber compliance. Test
`rate,
`bypass flow and stroke were kept constant. As can be
`clearly seen, valve loading forces exceed 50 N and,
`thus, are far above physiologic loading.
`
`
`
`Discussion
`
`This pilot study demonstrates that physiologic valve
`loading can be reproduced under appropriate acceler-
`ated testing conditions. However, testing under these
`conditions requires special testing compartments with
`an increased number of control parameters for the
`adjustment of physiologic loads. It also requires a two—
`step testing approach: first, physiologic loading forces
`have to be determined within a real-time circulatory
`mock loop; and second, this loading has to be repro-
`duced within the accelerated tester. For proper transfer
`of physiologic real-time loading forces to accelerated
`testing the test valves have to be mounted within a cal-
`
`PAGE 4 OF 7
`
`
`
`Durability/wear testing of heart valve substitutes
`H. Reul, K. Potthast
`
`155
`
`Force/N—-
`
`U)a:O0
`
`D Ap/mmHg—>
`
`Figure 8: Integral loading force and pressure difference across
`valve for a 27 mm SIM valve at a test rate of 1000 per min.
`
`Durability testing of bioprosthetic heart valves
`requires some additional discussion. The FDA and ISO
`guidelines for the wear testing of tissue valves are very
`similar to those for mechanical valves. Valves are to be
`
`tested for an equivalent of five years at a peak back-
`pressure of at least 90 mmHg for aortic and 120 mmHg
`for mitral valves. The valves should open and close
`completely. For stentless valves, the aortic wall should
`be modeled by a compliant tube.
`Unlike mechanical valves, tissue valves may show
`severe damage as a result of the wear test and may
`become dysfunctional before completion of the experi-
`ment. Possible damage to tissue valves include tissue
`tear, holes, delamination, abrasion, prolapse and stent
`fracture (5,8,9). Table I lists commonly observed dam-
`age in tissue valves, their cause, and critical test para-
`meters that influence the damage. Wear in mechanical
`valves is also included as a reference.
`
`Tissue tear is typically caused by high tensile stress—
`es. Finite element modeling (6) indicates that the high-
`est tensile stresses are in the region of the commissures
`150
`150
`
`27 mm SJM
`
`50 -100
`
`ApImmHg—>
`
`s o
`
`na
`
`o -
`
`Force/N—>
`
`0,00
`
`0.04
`
`0.08
`
`Figure 10: Integral loading force and pressure difference across
`valve for a 27 mm SIM valve at a test rate of 1000 per min,
`but at a preselected pressure difference of 120 mmHg (FDA—
`condition) across the valve.
`
`i Heart Valve Dis
`Vol. 7. No. 2
`March 1998
`
`
`
`
`
`
`Physiologic Range
`according to FDA-Guidelines
`
`70
`
`100
`so
`so
`Compliance Volume / ml—>
`
`110
`
`Figure 7: Integral loading force as a function of air compliance
`volume measured for a 27 mm SIM valve within the new test
`compartment at constant test rate (1000 per min), bypass
`throttle and stroke settings.
`
`ibration compartment, equipped with a force ring
`transducer. Once physiologic loading is adjusted and
`the parameter settings in terms of stroke volume, com-
`pliance volume and bypass flow are obtained,
`the
`valves are mounted Within the geometrically similar
`final test compartments and the continuously mea-
`sured pressure difference serves as the single control
`parameter for long-term studies.
`It could also be shown that at the back-pressures of
`the current standards the loading forces during accel-
`erated testing exceed the real-time loading forces by far
`and thus may provide unrealistically high valve loads.
`These initial findings indicate that corresponding
`amendments of the currently valid standards may be
`necessary.
`
`25 mm ISP
`
`1.00
`
`Force/N—>
`
`Physiologic Range
`
`Ap/mmHg—>
`
`0.0
`
`0,1
`
`0,3
`0,2
`Tlme / s ->
`
`0,4
`
`0.5
`
`Figure 9: Integral loading force and pressure difference across
`valve for a 25 mm ISP bioprosthesis at a test rate of 600 per
`min.
`
`PAGE 5 OF 7
`
`PAGE 5 OF 7
`
`
`
`156 Durability/wear testing of heart valve substitutes
`H. Real, K. Potthast
`
`I Heart Valve Dis
`Vol. 7. No. 2
`March 1998
`
`
`
`
`Table I: Wear and damage observed in cardiac valve prostheses daring accelerated wear testing.
`
`Valve type
`Wear / damage
`Cause
`Critical physical
`Critical test
`
`parameters
`parameter
`
`Mechanical g
`Sliding wear
`Sliding motion
`Range of motion
`Full opening and heart
`
`valve
`closing
`
`
`Impact wear
`Inertia at impact
`Impact force
`Ap, dp/dt, impact force
`
`Ap
`Pressure force at
`High tensile stresses
`Tissue tear at
`Stented tissue valve
`
`commissure closure
`
`Tissue abrasion at
`commissure
`during valve opening
`tissue and stent during
`
`during valve opening
`Tissue deflection
`Contact between
`Abrasion at the base
`AP
`
`at closure
`of the stent
`. leaflets and stent
`
`Contact between
`
`Tissue deflection
`
`Maximum opening
`
`Delamination of
`High bending
`tissue
`stresses, tissue
`during cardiac cycle
`
`buckling
`
`Tissue deflection
`
`Ap, maximum opening
`
`Stentless tissue valve
`
`Tissue tear at
`Ap, pm, aortic wall
`commissure
`closure, aortic wall
`compliance
`
`motion
`
`High tensile stresses
`
`Pressure force at
`
`Contact between
`Ap, pm, aortic wall
`leaflets and inflow
`closure, aortic wall
`compliance
`of cusp
`
`rim
`motion
`
`Abrasion at the base
`
`Leaflet deflection at
`
`Delamination of
`High bending stresses,
`Ap, pm, maximum
`tissue
`during cardiac cycle
`tissue buckling
`opening, aortic wall
`
`compliance
`
`Tissue deflection
`
`Ap, ptm,
`Axial and radial
`High tensile stresses
`Tear in aortic wall
`
`pressure loading
`wall compliance
`AP: back-pressure across the leaflets; ptm: transmural pressure across the aortic wall.
`
`in response to the pressure loading at closure. It is
`therefore important to control the closing pressure
`accurately. In the case of stentless valves, the stresses in
`the leaflets may be redistributed by the flexible aortic
`wall which may act as a shock absorber. Proper model-
`ing of the aortic wall compliance and control of the
`transmural pressure gradients should therefore be con-
`sidered when testing stentless valves.
`tissue
`Tissue abrasion is caused by rubbing of
`against the stent, stent cover or tissue reinforcement. It
`most likely occurs close to the commissures or at the
`base of the cusps where excursion of the tissue is max—
`imum during the cardiac cycle. Maximum deformation
`of the leaflets in the cusp region occurs at maximum
`back-pressure, while maximum deformation of the
`leaflets at the commissure occurs during full opening.
`Thus, control of both peak back-pressure and maxi-
`mum leaflet opening is important when testing for
`leaflet abrasion. It should be noted that the high cycle
`
`rate in accelerated wear testers may not create the same
`leaflet deflection or stresses as in a real—time pulse
`duplicator. Vesely et al. (10) pointed out that porcine
`tissue is viscoelastic, i.e. the stress-strain relationship is
`time-dependent.
`Delamination in the tissue is due to internal shear
`
`stresses. Vesely and Boughner (7) have shown that
`porcine tissue loses some of its natural ability to shear
`when tanned. As a result, bending of tanned tissue
`leads to high internal shear stresses, and potential
`buckling of the tissue (8). Delamination is typically
`observed in the hinge area of porcine valves where the
`tissue is relatively thick.
`As indicated in Table 1, critical test parameters for the
`testing of tissue valves include the maximum back-
`pressure and the maximum valve opening. In some
`commercially available durability testers,
`the back-
`pressure is controlled by a throttle in the flow loop that
`adjusts the bypass flow during closure. As pointed out
`
`PAGE 6 OF 7
`
`PAGE 6 OF 7
`
`
`
`Durability/wear testing of heart valve substitutes
`H. Reul, K. Potthast
`
`157
`
`References
`
`1. Reul H, Eichler M, Potthast K, Schmitz C, Rau G. In
`vitro testing of heart valve wear outside of the manu-
`facturer’s laboratory-requirements and controversies.
`I Heart Valve Dis 1996;5(Suppl. I):S97—SlO4
`2. Rau G, Reul H, Eichler M, Schreck S, Wieting DW. The
`effect of left ventricular dp/dt on the in vitro dynam-
`ics of Bjork—Shiley convexo-concave mitral valves. I
`Heart Valve Dis 1995;4(Supp1. I):Sl 7-820
`3. Schreck S, Inderbitzen R, Chien H, et al. Dynamics of
`Bjork—Shiley convexo-concave mitral valves in sheep.
`I Heart Valve Dis 1995;4(Suppl. I):S 21-825
`4. Knott E, Reul H, Knoch M, Steinseifer U, Rau G. In
`
`vitro comparison of aortic heart valve prostheses. I
`Thorac Cardiovasc Surg 1988;96(6):S952—S961
`5. Walley VM, Keon WI. Patterns of failure in Ionescu—
`Shiley bovine pericardial bioprosthetic valves. I Tho—
`rac Cardiovasc Surg 1987;913:5925-8933
`6. Krucinski S, Vesely I, Dokainish MA, Campbell G.
`Numerical simulation of leaflet flexure in bioprosthet-
`ic valves mounted in rigid and expansile stents. I Bio—
`mech 1993;26(8):S929-S943
`
`7. Vesely I, Boughner DR. Analysis of the bending
`behaviour of porcine xenograft leaflets and of natural
`aortic valve material. Bending stiffness, neutral axis
`and
`shear
`measurements.
`I
`Biomech
`1989;22(6/ 7):3655-5671
`
`8. Vesely I., Boughner DR, Song T, Tissue buckling as a
`mechanism of bioprosthetic valve failure. Ann Thorac
`Surg 1988;416:5302—8308
`9. Thiene G, Bortolotti U, Valente M, et al. Mode of fail—
`
`ure of the Hancock pericardial xenograft. Am I Car-
`diovasc Surg 1989;63:5129-8133
`10. Vesely I, Boughner DR, Leeson-Dietrich I. Biopros-
`thetic valve tissue viscoelasticity:
`Implications on
`accelerated pulse duplicator testing. Ann Thorac Surg
`1995;60(2):S379-S382
`
`I Heart Valve Dis
`Vol. 7. No. 2
`March 1998
`
`earlier, tissue valves may experience notable damage
`during the experiments which often leads to increased
`backflow leakage. To maintain the required peak back—
`pressure, the bypass resistance has to be increased to
`compensate for leakage through the test valves. It is
`therefore desirable to have individual throttles for each
`
`valve in order to maintain the same back—pressure for
`all valves throughout the duration of the experiment.
`The second important test parameter is the degree of
`opening of tissue valves. According to the regulatory
`guidelines, full opening should be achieved in the
`durability tester.
`’Full opening’
`is somewhat
`ill-
`defined, as the degree of opening is a function of the
`flow rate. At Baxter Laboratories, for instance, the
`
`valve is first tested in a pulse duplicator at a cardiac
`output of 5 1/ min. The degree of opening during peak
`forward flow is recorded with a Video camera posi-
`tioned along the axis of the valve. When the valve is
`placed in the durability tester, the stroke of the actuator
`is adjusted to reproduce the same degree of opening as
`in the pulse duplicator. Video images of the valve in
`both systems are used to verify proper tuning.
`In conclusion, the above discussion addresses some
`of the basic considerations for the testing of tissue
`valves. Because of the complex biomechanical proper—
`ties of tissue, it is unclear if and how the accelerated
`
`cycling rate is modifying the stress and strain distribu-
`tion in the valves (10). Furthermore, biochemical
`
`degradation and in situ host response are not consid—
`ered. Biochemical degradation or host overgrowth may
`modify the tissue properties and alter the motion and
`stress distribution in the tissue. Results of accelerated
`
`wear testing of tissue valves should therefore be
`reviewed in the context of the limitations of the test.
`
`Additional knowledge about the viscoelastic behavior
`of fresh and aged tissue will be necessary to further
`refine the accelerated wear testing of tissue valves.
`
`-
`Acknowledgments
`The authors thank Stefan G. Schreck PhD, Director,
`
`Engineering Research and Standards at Baxter Health-
`care Corporation, Irvine, CA and Carlos Rios from the
`same group for their most valuable contribution to the
`discussion on specific aspects for accelerated testing of
`bioprostheses.
`
`PAGE 7 OF 7
`
`: I
`
`. l ,
`
`PAGE 7 OF 7
`
`