`and durability of pericardial bioprostheses
`
`J FISHER, I J REECE, G R JACK, L CATHCART, D J WHEATLEY
`Department of Cardiac Surgery, University of Glasgow
`
`The design, function, and durability of four pericardial bioprostheses have been studied in the laboratory.
`Variations in design and construction affected both the pressure difference across the valves and the leaflet
`dynamics. In the durability tests tissue failure was found in all valves, with tears at the edge of the leaflets caused
`by abrasion on the cloth-covered inner frames. In the lonescu-Shiley standard valves, tears were also detected
`at the coaption sutures. These results have been compared with failure modes in explanted clinical valves.
`
`INTRODUCTION
`The Ionescu-Shiley (ISU) pericardial bioprosthesis has
`been used clinically for over 13 years with good clinical
`follow-up results (Ionescu et al., 1982). However, concern
`exists over the incidence of primary tissue failure report-
`ed clinically in these valves (Gabbay, et al., 1984a; Brais
`et al., 1985; Gallo et al., 1985). Tears have been found in
`leaflets of explanted valves close to the edge of the cloth-
`covered frames which have caused leaflet prolapse and
`regurgitation. Several different reasons have been sug-
`gested for tissue failure in ISU pericardial valves; abra-
`sion of the leaflets on the cloth-covered inner frame
`(Gabbay et al., 1984b); abrasion of the leaflet on the cloth
`buttress on the outside of the post (Martin et al., 1980);
`stress concentrations around the coaption stitches inside
`the top of the posts (Rainer, 1985); and bending and
`flexion stresses close to the edge of the support frame
`(Thubrikar et al., 1982; Ionescu et al., 1981). More
`recently, three new low profile pericardial bioprostheses,
`the Ionescu-Shiley Low Profile (ISLP), Hancock Peri-
`cardial (HP) and Mitral Medical (MM) valves have
`become available for clinical use in the United Kingdom.
`Our early clinical experience in Glasgow with two of
`these valves (ISLP and HP) has shown that primary
`tissue failure has not been eliminated in these new valve
`designs (Reece et al., 1986).
`In this study we have investigated the design, con-
`struction, function, and durability of four different peri-
`cardial valves (ISU, ISLP, HP, MM). The leaflet
`geometries and valve design have been analysed for the
`size 29 mm mitral valves, and the function of both size
`29 mm mitral and size 23 mm aortic valves has been
`assessed in our pulsatile flow test apparatus. Durability
`studies have also been carried out on size 29 mm valves.
`The results of these studies have given a clearer under-
`standing of the mechanisms of primary tissue failures
`found in explanted clinical valves.
`
`MATERIALS AND METHODS
`
`Description of the valves
`The key dimensions of the valves are given in Table 1.
`The external diameter is the diameter of the cloth-
`covered frames and leaflets, but excludes the flexible
`sewing ring, and the internal diameter defines the poten-
`tial orifice for forward flow. The leaflet height, h, defines
`the height of the flexing portion of the leaflet and h, the
`maximum depth of coaption between
`the closed
`unloaded leaflets.
`0 MEP Ltd 1987 Vol. 16, No. 2
`
`0046-2039/87/0004-0105 $02.00
`
`105
`
`Although the fundamental construction of all four
`valves is similar, with the leaflets sewn onto the outside
`of cloth-covered frames, there are significant differences
`in the detailed design and construction of these valves. In
`the ISU and ISLP valves the pericardial tissue appears
`to be initially fixed with glutaraldehyde in its natural
`shape and then mounted on the valve frame, while in the
`H P valve fresh tissue is mounted on the frame and fixed
`with glutaraldehyde under a small back pressure which
`defines the geometry of the unloaded closed leaflets. In
`the MM valve the leaflets are individually fixed on
`moulds which determine the shape of the unloaded
`flexing portion of the leaflet. The geometries of the
`frames and leaflets also differ (Table 1). The size 29 mm
`ISU and ISLP valves have an unloaded closed geometry
`defined by a cylindrical surface with radius R , , approx-
`imately 20 mm. This corresponds to a broad ‘U shape
`cut out in the base of the frame and a parallel portion at
`the top of each post which gives a slightly greater coap-
`tion depth, h,, between the leaflets. The height of the
`leaflets is reduced by 1 mm in the ISLP valve. The leaf-
`lets of the MM and H P valves are formed in approx-
`imately spherical shapes during fixation with a radius of
`curvature, R, approximately 15 mm. The overall leaflet
`height, h, is less; the posts have only a short parallel
`portion at the top and the coaption between the
`unloaded leaflets is less than in the ISU and ISLP valves.
`In all the valves the leaflets are sutured to the outside of
`the cloth-covered frames around the base of the valves
`and at the back of the posts. In addition, sutures are
`placed through the leaflets close to the top of the posts to
`ensure closure of the leaflets around the posts. In the ISU
`
`Table 1. Key dimensions in mm for the size 29 and
`size 23 mm valves
`ISU
`29
`31/30
`25
`20
`16
`4
`ISU
`23
`28/25
`19.5
`18.5
`
`Valve
`size
`
`Outside diameter, Do
`Internal diameter, Di
`Overall height, H
`Leaflet height, h
`Coavtion devth. h.
`Valve size
`
`Outside diameter, Do
`Internal diameter, Di
`Overall height, H
`
`ISLP HP MM
`29
`29
`29
`31/30
`29
`29
`24.5
`24
`24.5
`19
`17
`16.5
`15
`12.5
`12.5
`2.5
`2.5
`3
`ISLP HP MM
`23
`23
`23
`24
`22.5
`26/25
`19.5
`20
`19
`17
`15
`16
`
`Colibri Heart Valve LLC, Exhibit 2021, Page 1 of 5
`
`
`
`valves the coaption sutures are placed inside the posts
`while in the ISLP valves two sutures are placed through
`the leaflets and the side of the posts. In the HP valve
`sutures are placed through the leaflets above the posts
`and also through the side of the posts, while in the MM
`valve the suture is placed over and at the outside of the
`post so it is less effective in closing the leaflets together.
`The frames in the ISLP, HP, and MM valve are manu-
`factured from flexible polymers, while in the ISU it is
`manufactured from rigid titanium.
`
`Function tests
`Our pulsatile flow test apparatus has been described pre-
`viously (Fisher et al., 1986). Pressure difference and
`regurgitant volumes were measured for one size 29 mm
`and one size 23 mm valve of each type under a range of
`flow conditions A to E corresponding to cardiac outputs
`of 3-8.5 1 min-'. Leaflet dynamics were studied in the
`size 29 mm mitral valves under steady and pulsatile
`flows and movements recorded on video camera at
`20 mS intervals and with synchronised flash photogra-
`phy at 5 mS intervals.
`
`Durability tests
`One ISU, four ISLP, three H P and two MM valves, all
`size 29 mm were cycled in Rowan Ash accelerated
`fatigue testers at 12 Hz with a closed back pressure of
`between 100 and 130 mm Hg and peak forward flow of
`between 330 and 450 ml s-'. Valves were tested to
`failure, defined as a tear of at least 2-3 mm in one leaflet.
`
`Clinical experience
`Over the past four years 105 ISLP and 98 H P valves
`have been implanted in Glasgow with a mean follow-up
`time of 39 months per patient and 26 months per patient
`in each series. Primary tissue failure has occurred in six
`explanted ISLP valves and two explanted H P valves. In
`all cases the tears occurred in the leaflets close to the top
`of the cloth-covered posts causing leaflet prolapse and
`large regurgitation. A similar failure mode has been
`found in one explanted IS valve. The regurgitation in
`three of these explanted valves was measured in the pulse
`duplicator.
`
`RESULTS
`Measurements of mean pressure difference plotted
`against RMS forward flow are given in Fig. 1 for the size
`29 mm valves and in Fig. 2 for the size 23 mm valves. In
`all cases the orifice area of the open valve was deter-
`mined by the opening of the free edges of the leaflets.
`This opening was widest in the MM 29 mm valve which
`had the lowest pressure difference and smallest in the
`ISU 29 mm valve which had the highest pressure gra-
`dient. The opening of the free edges of the leaflets was
`dependent on the position of the coaption stitches at the
`top of the posts. The size 23 mm MM and H P valves had
`lower pressure drops than the size 23 mm ISLP and IS
`valves. The regurgitant volumes are given in Fig. 3 for
`the size 29 mm valves and Fig. 4 for the size 23 mm
`valves. There was little difference in the regurgitation
`between the different valves. The closing volumes in the
`size 29 mm valves were larger than in the size 23 mm as
`the volume swept back by the leaflets was greater. The
`closed regurgitation occurred through the cloth sewing
`ring on the outside of the valve frame and was greater in
`
`106
`
`0.8
`
`0
`-c
`3.4
`
`1
`
`/. - 1 2
`
`ISLP 29
`
`ISU29
`
`lo-
`
`/ 1:
`
`Y 4
`VI
`
`0,
`c 8 2
`
`E
`
`J*
`
`./*
`
`2-
`
`- 0 8
`
`/' -0 4
`1.
`*.P
`
`m
`P *
`
`0
`
`160
`
`260 360 460 560
`
`o-
`
`360
`
`460
`
`560
`
`i
`
`1.
`
`i
`
`l
`
`
`
`-
`
`Fig. 2. Mean pressure diflerence plotted against rms .forward
`flow for the size 23 mm valves
`
`closed
`D closing
`
`MM 29
`ISU 29
`I S L P 2 9
`HP 29
`Fig. 3. Regurgitant volumes for the size 29 mm valves
`Engineering in Medicine 0 MEP Ltd 1987
`
`Colibri Heart Valve LLC, Exhibit 2021, Page 2 of 5
`
`
`
`-
`
`closed
`closing
`
`z 1
`e -
`
`c
`I ?
`m ?
`
`HP 23
`ISU 23
`ISLP23
`MM 23
`Fig. 4. Regurgitant volumes for the size 23 mm valves
`
`the size 23 mm valves due to the longer diastolic time
`interval in the aortic position and a different type sewing
`ring.
`Studies of the leaflet dynamics in the size 29 mm valves
`showed that the free edge symmetrical orifice in the ISU
`and H P valves due to the restricted opening and tension
`induced in the tissue by the coaption stitches, whereas in
`the ISLP and MM valves the tissue at the free edge was
`under less tension and the orifice was not as uniform. In
`the open position the leaflets of the IS and ISLP formed
`a straight line at the edge of the frame, while in the H P
`and M M valves the tissue which was formed in a small
`radius over the edge of the frame during fixation formed
`an 'S' configuration when the leaflets were fully open
`(Fig. 5). In the closed position the posts of the flexible
`frames deflected inwards reducing the tension at the free
`edge of the leaflets. At low steady flows the ISLP and
`ISU valves opened fully at 40 ml s-', the H P valve at
`80 ml and the MM valve at 120 ml s-'. All valve leaflets
`opened at the lowest pulsatile flow (A) corresponding to
`a peak flow of 150 ml s-'. The methods of fixation also
`affected the way the valve leaflets transposed from the
`closed to the open position. In the ISU and ISLP valves
`the tissue buckled circumferentially across the leaflets,
`while in the H P and MM valves the tissue reversed its
`curvature in the base of the leaflet, first buckling in the
`radial direction. Closure was the reverse of opening in
`each valve.
`All but one of the valves failed prematurely after less
`than 60 million cycles with tears in the leaflets at the edge
`
`Leaflet
`
`.Leaf let
`
`-Ste mnt
`
`,Stent
`
`H P
`ISU
`MM
`ISLP
`Fig. 5. A vertical section through the base of the leaflets at the
`edge of
`the open
`the frames showing the geometry of
`leaflets
`0 MEP Ltd 1987 Vol. 16, No. 2
`
`Fig. 6. Results of the accelerated fatigue tests
`
`of the cloth-covered frames (Fig. 6). The other valve
`failed after 320 million cycles in a similar manner. All the
`leaflets of the failed valves showed excessive thinning and
`abrasion adjacent to the edge of the cloth-covered
`frames. This was greatest about half way up the posts at
`the shoulder of the scallop which was the origin of the
`tears in most of the valves. In some valves these tears
`extended to the free edge of the leaflet. In one ISLP and
`one H P valve the tears started closer to the top of the
`posts. In all the HP, ISLP, and MM valves the tissue was
`intact and in good condition around the coaption
`sutures at the top of the posts. In the IS valve small tears
`occurred in two leaflets at the coaption sutures, although
`the largest tear originated from half way up the post
`
`Fig. 7. Failed ISU valve from the fatigue tests
`
`107
`
`Colibri Heart Valve LLC, Exhibit 2021, Page 3 of 5
`
`
`
`Fig. 8. A failed ISLP valve from the fatigue tests
`
`(Fig. 7). Figures 8-10 show examples of the failed ISLP,
`HP, and MM valves.
`Figure 11 shows an example of an explanted ISLP
`valve with torn leaflet, and Fig. 12 shows a failed
`explanted HP valve. In both cases the tears in the leaflets
`occurred at the top of the posts. It was difficult to deter-
`mine whether the tears originated from the edge of the
`cloth-covered post or from the coaption sutures passing
`through the side of the posts. Abrasion of the leaflets
`adjacent to the frame was much less than in the valves
`cycled in the fatigue tester and the leaflets were rein-
`forced with host tissue ingrowth over these areas. Tissue
`ingrowth was less towards the top of the posts. An
`explanted ISU valve with a torn leaflet is shown in
`Fig. 13. Tears appeared both at the top of the posts and
`at the shoulder of the scallop, and a small tear was seen
`at the coaption stitch in one of the other leaflets. The
`regurgitation measured in three explanted valves with
`torn leaflets was between 40 and 60 per cent.
`
`Fig. 10. A failed M M valve from the,fatigue tests
`
`DISCUSSION
`Although we have shown significant differences in the
`design and construction of the four pericardial valves
`studied, there was little variation in measurements of
`pressure drop and regurgitation between the valves. The
`coaption sutures at the top of the posts did restrict the
`opening of the valve leaflets particularly in the ISU valve
`which had the largest pressure difference of the size
`29 mm valves. The different design and methods of con-
`struction had a greater effect on the leaflet dynamics.
`Both the closed and open position of the leaflets, the
`movements between the two positions, and the bending
`stresses at the edge of the open leaflets were affected by
`the leaflet geometries and methods of tissue fixation. In
`addition, the positioning of the coaption suture affected
`the tension in the free edge of the open leaflet and the
`flexibility of the frames affected the tension in the closed
`leaflet. These different leaflet dynamics, however, did not
`affect the results of our durability tests.
`
`Fig. 9. A failed H P valve from the fatigue tests
`
`108
`
`Fig. 11. An explanted ISLP valve with a torn leaflet
`Engineering in Medicine $1 MEP Ltd 1987
`
`Colibri Heart Valve LLC, Exhibit 2021, Page 4 of 5
`
`
`
`All the valves in the fatigue tests failed due to abrasion
`and thinning of the leaflets as they were stretched over
`the edge of the cloth-covered frames in the closed posi-
`tion. The abrasion was greatest half way up the posts and
`this was the origin of the tears in most of the valves. Only
`in the ISU valve where the coaption stitch was placed
`inside the post did the tears originate at the coaption
`stitches. The accelerated fatigue tests can be considered
`artificially harsh for the pericardial valves with cloth-
`covered frames as biological effects such as tissue
`ingrowth and blood deposits on the cloth, which reduce
`the abrasion in vivo, are not simulated in the tests.
`Explanted clinical valves showed similar tears to the
`fatigue tested valves but the tears were usually closer to
`the top of the posts where the tissue ingrowth was not as
`great. It is likely that abrasion of the leaflets on the cloth-
`covered posts also caused the tears clinically in the ISLP
`and HP valves. Tears were only clearly detected at the
`coaption stitches in the explanted ISU valve, which cor-
`relates with the fatigue test findings. However, the coap-
`tion sutures could weaken the leaflets at the top of the
`posts in the ISLP and H P valves as they are positioned
`close to an area that is being thinned by abrasion. This
`area is also under high bending stresses in the open posi-
`tion.
`Only after the abrasion to the leaflet at the edge of the
`frame has been eliminated can the effect of the different
`leaflet dynamics on valve durability be assessed.
`
`REFERENCES
`Brais, M. P. et al. (1985) Ionescu-Shiley pericardial xenograft follow-up
`of up to six years, Ann. Thorac. Surg., 39,105-1 11.
`Fisher, J. et al. (1986) Design of a function test apparatus for prosthetic
`heart valves. Initial results in the mitral position, Clin. Phys. Physiol.
`Meas, 6,63-73.
`Gabbay, S. et al. (1984a) Longterm follow-up of the Ionescu-Shiley
`mitral pericardial xenograft, J . Thorac. Cardiovasc. Surg., 88,758-763.
`Gabbay, S. et al. (1984b) Fatigue induced failure of the Ionescu-Shiley
`pericardial xenograft in the mitral position, J . Thorac. Cardiooasc.
`Surg., 87,836-844.
`Gallo, 1. et al. (1985) Incidence of primary tissue failure with the Ion-
`escu-Shiley pericardial valve, J . Thorac. Cardiovasc. Surg., 90,278-280.
`Ionescu, M. I. et al. (1981) Low profile prosthetic heart valve, European
`Patent Application 813051 12224.
`Ionescu, M. I. et al. (1982) Clinical experience of the pericardial xeno-
`graft, 11 years experience, in Cardiac Bioprostheses, (Edited by Cohn, L.
`H. and Gallucci, V.) Yorke Medical Books, New York, pp. 42-60.
`Martin, T. R. P. et al. (1980) Accelerated fatigue tests on biological
`tissue valves, Proc. ESAO, 7,315-318.
`Rainer, W. G. (1985) Discussion of in vivo haemodynamic comparison
`of porcine and pericardial valves, J . Thorac. Cardiovasc. Surg., 89,
`358-368.
`Reece, I. J. et al. (1986) A comparison of experience with the Hancock
`and Ionescu-Shiley low profile pericardial bioprosthesis, 2. Kardiol.,
`75 (SUPPI.), 232-236.
`Thubrikar, M. J. et at. (1982) Design and stress analysis of bioprosthetic
`valves in vivo, in Cardiac Bioprostheses, (Edited by Cohn, L. H. and
`Gallucci, V.) Yorke Medical Books, New York, pp. 445-455.
`
`Fig. 12. An explanted H P valve with a torn leaflet
`
`Fig. 13. An explanted ISU valve with a torn leaflet
`
`0 MEP Ltd 1987 Vol. 16, No. 2
`
`109
`
`Colibri Heart Valve LLC, Exhibit 2021, Page 5 of 5
`
`