- Email: [email protected]
Analysis of the Medtronic Intact bioprosthetic valve *
J
THORAC CARDIOVASC SURG
1991;101:90-9
Analysis of the Medtronic Intact bioprosthetic valve * Effects of "zero-pressure" fixation The long-term performance of current-design porcine xenograft valves bas not been satisfactory. These valves are generally fixed at "low pressures" of about 3 to 5 mm Hg. The Medtronic Intact (Medtronic, Inc., Minneapolis, Minn.) valve is fixed at "zero pressure" and is proposed as a better alternative to existing xenograft valves. A mechanical analysis of this valve bas been carried out to determine if the Intact valve differs significantly from the low-pressure fixed xenograft Twelve circumferential strips of tissue 5 nun wide were cut from the leaflets of four clinical-grade Intact valves. Their stress/strain, stress relaxation, and flexural bebavior were examined mechanically and histologically. The Intact valve was more extensible tban the low-pressure fixed xenograft (22 % versus 12 % strain, P < 0.001~ relaxed faster (p < 0.001~ and was more pliable tban the xenograft (p < 0.05). It did not, bowever, buckle less tban did the low-pressure fixed xenograft during enforced bending, and it buckled significantly more tban did fresb porcine aortic valve tissue (p < 0.001). The Intact valve also relaxed significantly more slowly tban did the fresb tissue (p < 0.05). Its bending stiffness bad a stronger dependence on leaflet thickness tban the bending thickness of fresb tissue bad (p < 0.001) but a weaker dependence tban the bending thickness of the low-pressure fixed xenograft material bad (p < 0.001). The Intact valve demonstrated a very large variabilitiy in extensibility, bending stiffness, and buckling bebavior, witb little correlation between these parameters. Some valves appeared to bave wrinkled leaflets; others were likely fixed at different pressures. The shrinkage of the leaflet material at these low fixation pressures is likely important, since it can modify the elastic bebavior of the valve cusps. Overall, the Intact valve bad a more ''natural'' elastic bebavior tban bad low-pressure fixed xenograft, and it should therefore experience lower stresses during normal valve function. It can be concluded tbat zero-pressure fixation does preserve many of tbe desirable stress-reducing properties of aortic valve tissue.
Ivan Vesely, PhD, London. Ontario. Canada
Long-term clinical studies!" of bioprosthetic valve implants have demonstrated that the major deficiency of these devices is their poor durability. Although 95% still function after 5 years, as few as 40% remain functional From The John P. Robarts Research Institute, London, Ontario, Canada. Supported by a fellowship from the Heart and Stroke Foundation of Canada. Received for publication July 24, 1989. Accepted for publication Feb. 2, 1990. Address for reprints: Ivan Vesely, PhD, The John P. Robarts Research Institute, P.O. Box SOlS, London, Ontario, Canada, N6A 5K8. *Medtronic, Inc., Minneapolis, Minn.
12/1/20222
90
after 15 years.' The degenerative failure that produces this attrition results from the tearing and calcification of the valve leaflet material.' The exact mechanism of tissue degeneration remains unclear, but there is strong evidence that mechanical factors, particularly flexural fatigue, play an important role both in leaflet tearing and in calcification. 5-7 Previous analyses of porcine aortic valve leaflets have shown that current "low-pressure" glutaraldehyde-fixation techniques decrease tissue extensibility':" and shearing.!" and increase bending rigidity'? and the propensity to collapse and buckle during leaflet bending. I J The lowering of transvalvular pressures during glutaraldehyde fixation has been shown to preserve more of the natural extensibility and pliability of the leaflet materi-
Volume 101 Number 1
Analysis of Medtronic Intact valve 9 I
January 1991
400
300
FRESH PORCINE AORTIC VALVE TISSUE
CRYOPRESERVED AORTIC ALLOGRAFT
0
o, .x: c en
"'
200
QJ ....
en
100
o +---'~~--r--;;;;;";;;~::"'-~~----.-~~-.---l o 40 30 10 20
10
0
20
30
40
20
30
40
400 - r - - - - - - - - - - - - - - - - - - ,
o
LOW PRESSURE FIXED XENOGRAFT
300
o, .x: c
INTACT VALVE
::l 200 ~
en 100
10
20
10
30
Strain in %
Strain in %
Fig. I. Summary of stress/strain curves for eryopreserved aortic allograft tissue, fresh porcine aortic valve tissue, low-pressure fixed porcine xenograft material, and the Intact valve. Measurements for the first three were obtained previously!" and are included for comparison. Note the lower extensibility of the low-pressure fixed xenograft.
a1.12 It has even been suggested that open-position fixation produces a graft that functions better and may therefore last longer.' J Although most valve manufacturers have responded by lowering fixation pressures, the 3 to 5 mm Hg pressure they use to produce well-shaped cusps remains sufficiently high to significantly reduce leaflet extensibility and pliability.!? 14 Medtronic, Inc. (Minneapolis, Minn.) has recently introduced the Intact (Xenotech) porcine bioprosthesis, a valve that claims "zero-pressure" fixation. Fixing valves at zero pressure has been suggested as a better alternative to low-pressure fixation, since much of the collagen crimp can be preserved. I 2 This material should therefore demonstrate much of the pliability and extensibility of natural aortic valve leaflets, experience lower stresses during valve function, and hence last longer. A mechanical analysis of the Intact valve is therefore warranted to determine how well the mechanical behavior of the leaflet material has been preserved with zero-pressure fixation.
Methods Data acquisition. Twelve strips of tissue were obtained from the leaflets of four clinical-grade Intact porcine bioprosthetic valves.Tensile stress/strain and stress relaxation tests were performed with an Instron tensile testing machine (model 1125, Instron Corp., Canton, Mass.). The tissue was repeatedly stretched at 10 rum/min to a maximal load of 0.98 N (100 gm) to precondition the material and produce reproducible load/ elongation curves.l'' The final load/elongation curves were digitized and transformed to stress/strain curves to normalize for variable length and thickness. Both load and tissue elongation were obtained directly from the chart paper during digitizing on a Summagraphics palette (model MM 1812, Fairfield, Conn.). The errors that may have occurred during plotting and digitizing were therefore negligible compared with those potentially introduced during measurement of specimen width, length, and thickness. The precision in cutting the tissue strips to width was estimated to be ±0.2 mm (4% for as mm wide strip). Thickness was measured with a special gauge constructed for this purpose." and gauge length was measured with a calibrated scale affixed to the tissue-clamping mechanism. The precision of the thickness measurements was estimated to be ± 0.05 mm and that of the gauge length to be ± 0.2 mm. The tissue strips were clamped between machined stainless-steel jaws that were lined
92
The Journal of Thoracic and Cardiovascular Surgery
Vesely
100 90 0\ C
'c '0
E
80
Q)
L-
en en
.... Q)
70
L-
ill1
FRESH PORCINE AORTIC VALVE TISSUE
CRYOPRESERVED AORTIC ALLOGRAFT
(f)
60 50 0.1 100
1.0
10.0
100.0
1000.0 0.1
-
1.0
10.0
100.0
1000.0
10.0
100.0
1000.0
90 0\
c
'c '0
E Q)
80
L-
en en
.... Q)
70
L-
(f)
ill1
60 50 0.1
LOW PRESSURE FIXED XENOGRAFT
1.0
10.0
INTACT VALVE
100.0
1000.0 0.1
TIme in seconds
1.0
Time in seconds
Fig. 2. Summary of stress relaxation curves for cryopreserved aortic allograft tissue, fresh porcine aortic valve tissue, low-pressure fixed porcine xenograft material, and the Intact valve. Measurements for the first three were obtained previously!" and are included for comparison. Note the lower relaxation slopes of the two glutaraldehyde- . fixed tissues.
with No. 320 emery cloth to prevent slippage. Admittedly, tensile testing of cut and clamped tissue strips can introduce certain procedural errors. Tensile testing of this type, however, has become an accepted standard, and it is used routinely to assess the mechanical properties of various bioprosthetic tissues" 16-18 It is particularly valid when different tissue types are compared under identical conditions in the same laboratory. For stress relaxation testing, the specimens were held elongated and the.induced tensile force was recorded over a 5-minute interval. Since previous experience with aortic valve tissue has shown that relaxation rates do not vary with applied stress, we did not control for maximum stress level. Each specimen was allowed to relax from a constant load of 0.98 N. All strips were 5 mm wide, cut in the circumferential direction, and tested in a physiological saline bath heated to 37° C. All stress calculations are based on "engineering stress," not "true stress," using the original tissue thickness instead of accounting for thinning during elongation. This approximation avoids estimating the Poisson ratio and the lateral contraction of the specimen during the course of the test. Both of these estimates are required to closely approximate the exact specimen cross-sectional area at any given strain. A more detailed description of the testing procedure can be found in the report by Vesely and colleagues."
Bending tests were performed on the same 12 strips using a custom-built tissue-bending machine. 19 In a heated saline bath, the strips were bent through a 9Q-degreearc to varied curvatures and their bending moments measured to an accuracy of I fJ.N . m (0.1 gm . mm). Each strip was tested four times to establish repeatability." Following bending tests, the strips were cut in halflengthwise, held bent with surgical clips, histologically processed, sectioned to a thickness of 7 fJ.m, stained with van Geison's stain, and examined with low-power light microscopy to measure the degree to which the tissue buckled during enforced bending. I I Data for all tests were compared with data for fresh porcine aortic valve tissue obtained from a local abattoir, low-pressure fixed porcine xenograft tissue obtained from St. Jude/Biolmplant (St. Hyacinthe, Quebec, Canada), and cryopreserved human aortic valve tissue obtained from the Deborah Heart and Lung Center (Browns Mills, N.J.). Data for these tissues had been acquired and published previously." Data analysis. The elastic modulus of each strip was calculated froin the stress/strain curve at a stress of 300 kPa. This stress level simulated the tension the leaflets would experience during diastolic valve closure.l'' The length of the "low-modulus phase" of each stress/strain curve, the leaflet extensibility,
Volume 101 Number 1
Analysis of Medtronic Intact valve
January 1991
93
40,--------------------,
o
N
E
!
CRYOPRESERVED AORTIC ALLOGRAFT
A INTACT VALVE
30
ff) ff)
w
i'u,f.
20
ti (:J
~ 10 z
W
rn O+--t__---+--+--+-+--+-~t____+-_+_-__1
A 0.0
0.1
0.2
0.3
0.4
0.5 0.6 (mm)
0.7
0.8
0.9
1.0
THICKNESS
40,-----------------------,.,-,
40 N
•
E
FRESH PORCINE AORTIC VALVE TISSUE
z 30 --S
A INTACT VALVE
ff) ff)
.. . _-e:-.. --" ---.--
w
zu,
u, ~
20
ff)
•
z 10 z w
rn
a
6.
Z 30 --S
.... INTACT VALVE
E
0.1
0.2
0.3
0.4
0.5 0.6 THICKNESS (mm)
LOW PRESSURE FIXED XENOGRAFT
ff) ff)
w
i'u,f.
T
ff)
(:J
1
0.7
20
~
A
_ .:.---
(:J
0
B 0.0
N
~ 10
z
W
m
0.8
0.9
1.0
c
a +--+--+---f''-----i''''=-t--+----l--+--f----J 0.0
0.1
0.2
0.3
0.4
0.5 0.6 (mm)
0.7
0.8
0.9
1.0
THICKNESS
Fig. 3. Bending stiffness plots of the Intact valve compared with the cryopreserved aortic allograft (A),fresh porcine aortic valve tissue (B), and low-pressure fixed porcine xenograft material (C). The data points and error bars represent the mean and standard deviation of bending-stiffness measurements obtained from four bending curves for each tissue strip. The lines of best fit were obtained by least-squares regression and demonstrate the dependence of bending stiffness on leaflet thickness.
Table I. Calculated values from the curves in Figs. 1 and 2 Elastic modulus at a stress of 300 kPa (MPa)
Summary of tensile data (mean ± SD) I. Allograft 9.05 ± 5.35 2. Fresh tissue 13.01 ± 1.74 3. Xenograft 12.53 ± 3.02 4. Intact valve 9.68 ± 1.64 Comparison of groups* I vs 4 NSt 2 vs 4 p ~ 0.0025 3 vs 4 P ~ 0.05
Length of lowmodulus phase (%) 23.4 22.2 12.4 21.4
± ± ± ±
2.07 3.31 0.87 5.51
NS NS p
~
Slope of stress relaxation curve -5.9 -6.9 -3.5 -4.9
± 2.9
± 1.8 ± 0.8 ± 1.0
NS p
0.0001
~
0.05
p~O.OI
*(iroups "ere statistically compared with a small-sample Student's t test. t:\ot sig nilicaru.
was determined from the intersection of the x axis with a tangent to the stress/strain curve at the 300 kPa point. This parameter indicates the degree to which collagen crimp has been affected by the treatment technique. For example, when porcine valves are fixed at higher pressures, collagen crimp is eliminated and the tissue becomes less extensible. The slopes of the stress relaxation curves were calculated to evaluate integrity of the mucopolysaccharide matrix in each case. A steep slope indi-
cated rapid relaxation of the material through the realignment or migration of fibers in response to applied stress. Such relaxation is possible only if the ground substance is preserved during fixation. Slow relaxation indicates that the material behaves more like an elastic solid, which suggests that the fibers have become immobilized. Measurements of bending moment versus curvature produced bending curves, the slopes of which gave the bending stiffness. Measurements of bending stiffness were first
94
The Journal of Thoracic and Cardiovascular Surgery
Vesely
Fig. 4. Low-power photomicrograph of a typical thin section of buckled Intact valve tissue. Buckling depth was defined as distance (d) as shown, and normalized for the thickness of the specimen (t) to give fractional buckling depth. Buckling depth and tissue thickness were measured with a light microscope and calibrated eyepiece, and the radius of curvature (R) was measured directly from the mounted specimens by means of a hand-held optical comparator with calibrated arcs engraved on the objective. 10 Marker represents 0.1 mm
compared with analysis of variance. The bending stiffness was then plotted against the thickness of the tissue strip, and the slopes of the least-squares regressions of these data were compared by means of multiple Students's t tests. The extent of buckling was assessed by examining the thin sections with lowpower microscopy to measure buckling depth as described previously. I I The radius of curvature, tissue thickness, and buckling depth were measured at several areas around the bend for each of the two pieces cut from the Intact valve test strip. Buckling behavior was expressed as fractional buckling depth versus thickness X curvature, since an increase in both leaflet thickness and in bending curvature yields a corresponding increase in thedegrcc of buckling. I I The lines of best fit through the data sets were compared for elevation as described by Zar. 20 Some errors could have resulted from our methods. The Intact valves were tested for bending stiffness approximately I year after the other material was tested. Although it is conceivable that our bending machine may have produced slightly different results after this period of time, we have not found significant differences between identical fixed tissue tested in 1987 and at present, which makes this an unlikely source of error. Our technique of measuring the gauge length is capable of overestimating tissue extensibility. All specimens, however, were analyzed with the same technique. Our results are therefore internally consistent, and comparisons based on these results are valid. It should be noted, however, that the high variability in material-testing protocols makes the comparison of published studies difficult. 16 Admittedly, measurement of buckling depth
involved some subjectivity. The technique, however, was standardized to minimize this subjective influence, 11 and all tissue was examined by the same individual to control for this influence. Comparisons based on these measurements are therefore valid.
Results Tensile tests. The stress/strain and stress relaxation curves are shown in Figs. 1 and 2 for comparison of the Intact valve, low-pressure fixed xenograft, cryopreserved aortic allograft, and fresh pig aortic valve tissue. From Fig. I, it can be seen immediately that the low-pressure fixed xenograft is less extensible than the other three and that the Intact valve has a rather large spread in the extensibilities of individual samples. For detailed comparison, Table I shows a summary of the calculated parameters obtained from the tensile tests. Cryopreserved aortic allografts and fresh pig aortic valve tissue had similar elastic moduli at high stress, a similarly extensibility (length oflow-modulus phase), and nearly identical rates of stress relaxation. There was no significant difference between these two groups of unfixed tissues. Although the low-pressure fixed porcine xenograft leaflets had the same elastic modulus, they were much less extensible than
Volume 101
Analysis of Medtronic Intact valve
Number 1 January 1991
U.b , - - - - - - - - - - - - -
~
0.5
(:>
0.4
w
c: •
c:
~ ~
-----
0.3
U
OJ OJ ...J
-c z a
0.2
Ci
0.1 {,.{,.
;:: u
"-
A
0.0
---,
LOW PRESSURE FIXED XENOGRAFT FRESH PORCINE AORTIC VALVE TISSUE
o z
95
-@':-~/~:/;~~-
li. ....- . {,. A--
----
.......
-. - ...
•_e_---.
{,.
-.
_ .... _ - - ;
.---::
+--+-4IHt--+-----~>___jf______<~~~__+~-----J
0.5
0.0
1.0
1.5
THICKNESS x CURVATURE
0.6....------------------,
0.6 I f-
Cl.
w
0
(:>
z
0.5
•
FRESH PORCINE AORTIC VALVE TISSUE
•
INTACT VALVE
I f-
Cl.
u
OJ OJ ...J
« z
0.4
(:>
z
u -c a:: u,
0.4
~
u
0.3
OJ OJ
6t _.. . _.. . A
...J
0.2
;:: u
0.1
-c a::
u,
0.0
B 0.0
0.5
1.0
1.5
THICKNESS x CURVATURE
0.0
C
•
6.
"6-.-.. . l::J.
...__ ~_\.o'"
•
- --
--;:------
c:
~ _----
D. A li.
..._ ....::;....~~
{,..
0
0.1
..
. ... ----......---.... • IJ;,.
0.3
-c z
0.2
INTACT VALVE li.
~
0
;::
•
W 0
~ ~
{,. LOW PRESSURE FIXED XENOGRAFT
0.5
a--{:, {,.
• +-.........------..~__+_~-~__+_~~~-_+_-+---l
0.0
0.5
1.0
1.5
THICKNESS x CURVATURE
Fig. 5. Plots of fractional buckling depth (depth/tissue thickness) versus thickness X curvature for low-pressure fixed xenograft material compared with fresh tissue (A),and the Intact valve compared with fresh tissue (B)and with the xenograft (C). For the Intact valve, up to four buckling measurements were obtained from a single leaflet. The fresh and low-pressure fixed valve tissue was tested previously and is included only for comparison. I I
either the fresh porcine or the allograft tissue (12% versus 22%). This shows that even low-pressure glutaraldehyde fixation reduces tissue extensibility. The cross-linkage "freezes" the normally wavy fibers in their straightened configuration, and subsequent loading of the fixed material produces lower extensions. Fig. 2 and comparisons in Table I show that the fixed xenograft valves did not relax as quickly as the other three. This suggests that the mucopolysaccharide matrix has been either partially removed or cross-linked, so that a rapid reconfiguration of fiber geometry in response to stress cannot occur. The Intact valve had a significantly lower elastic modulus than those of the fresh and low-pressure fixed porcine valves and was more extensible than other xenograft tissue (Table I). Its extensibility was very similar to that of the fresh tissue. This would suggest that zero-pressure fixation does indeed preserve the natural extensibility of the aortic leaflet by fixing the collagen fibers in their naturally wavy configuration. Although the Intact valve relaxed more quickly than did the low-pressure fixed xenograft (relaxation slope of -4.9 versus -3.5,
p < 0.01), it relaxed more slowly than did the fresh tissue (slope of -4.9 versus -6.9,p < 0.05). The viscous behavior of the Intact valve is therefore better than that of conventional xenografts but still not as good as that of natural aortic valve tissue. It is interesting to note that the relaxation of the Intact valve was not significantly different from that of the cryopreserved allograft. It is likely that storage and freezing of the allograft valves facilitates the loss of some ground substance, just as does glutaraldehyde fixation. Bending tests. The results of the bending-stiffness measurements are shown in Fig. 3. Analysis of variance and a N ewman-Keuls test show no differences in bending stiffness among the Intact valve, the cryopreserved allograft, and the fresh pig valve tissue. The low-pressure fixed xenograft, however, was significantly stiffer than the others (p < 0.05). Zero-pressure fixation therefore produces a valve leaflet with a pliability comparable to that of unfixed tissue and superior to that of low-pressure fixation. It is interesting to note that thin Intact valve leaflets are more pliable than unfixed tissue but thicker leaflets are stiffer than unfixed tissue. This dependence of
The Journal of
96
Vesely
Thoracic and Cardiovascular Surgery
400.,-------------------,
300
INTACT # 2
o .x:
0.. C
.;;; 200 (J)
~
Vi 100
o +-.--.........~=;:::;:::;::::;=....~~-,---.-~..-.----1 o 40 10 30 20
0
10
20
30
40
400.,------------------,
300 o 0.. .x: c
INTACT # 3
INTACT # 4
';;;200 (J)
~
Vi
100
10
20 Strain in %
30
10
20 Strain in %
30
40
Fig. 6. Stress/strain curves for the four Intact valves tested. Note the presence of both intervalve and intravalve variability of leaflet extensibilities. Although the leaflets of Intact No.2 and Intact No.4 have similar extensibilities, . the two valves appear to have been fixed at different pressures. The high variability within Intact No. I and Intact No.3 probably results from wrinkled leaflets.
bending stiffness on leaflet thickness was analyzed by obtaining least-squares regressions for the four groups of bending-stiffness data and comparing them with multiple t tests. The Intact valve had a significantly greater slope (p < 0.(01) than had both the cryopreserved allograft (Fig. 3, A) and the fresh porcine aortic valve (Fig. 3, B). This slope, however, was lower (p < 0.001) than that for the low-pressure fixed xenograft (Fig. 3, C). Fixed tissue therefore demonstrated a greater increase of bending stiffness for a given change in thickness than did unfixed tissue, and this relationship intensified with fixation pressure. It should be noted that there is only a 0.006 probability of type I error with these multiple t tests. 20 Bending stiffness also depends both on the degree of cross-linkage between fiber layers and on the extensibility of the tissue. Cross-linkage increases bending stiffness by preventing shearing between fiber layers.!? and a reduced extensibility means that the material operates at a high modulus, even at the low strains occurring during
bending. A reduction in ability to shear also makes the material behave more like an elastic solid, with a bending stiffness strongly dependent on the sectional modulus (the second moment of area)." In such an "ideally elastic" material, the bending stiffness increases with the cube of the thickness, yielding a very high dependence of bending stiffness on leaflet thickness.'? The low glutaraldehyde concentrations used to fix the Intact valve may well have better preserved the ability of the valve to shear during bending, hence increasing its pliability. Buckling tests. A typical thin section of buckled Intact valve tissue is shown in Fig. 4. The analyses of such histologically prepared sections are shown in Fig. 5. Fig. 5, A, is a comparison plot of the buckling behavior of lowpressure fixed xenograft tissue and fresh porcine aortic valve tissue (obtained previously). II The difference in elevation of the two regression lines demonstrates that the xenograft buckled more than did the fresh tissue at all curvatures. This difference was significant (p < 0.(01).
Volume 101 Number 1 January 1991
Fig. 5, Band C, compares the Intact valve with the fresh tissue and with the low-pressure fixed xenograft. The Intact valve buckled significantly more (p < 0.001) than did the fresh tissue but not significantly less than did the low-pressure fixed xenograft. Like the xenograft, the Intact valve buckled at curvatures too low to produce any buckling in the fresh tissue. Compressive buckling induced in the leaflet during systolic valve opening has been advanced as a probable mechanism for bioprosthetic tissue disruption and cuspal tearing in vivo.11 The similar buckling behaviors of the Intact and the low-pressure fixed xenograft suggest that implanted Intact valves may also fail in this manner. Variability of measured parameters. The stress/ strain curves in Fig. 1 show that the extensibility of the Intact valve is considerably more variable than that of the others. To examine this variability further, the stress/ strain curves for each valve were plotted separately (Fig. 6). These plots show both intervalve and intravalve variability in the extensibilities ofthe cuspal tissue. Intact No. 2 and Intact No.4 are internally consistent but have different overall extensibilities (p < 0.(025). Conversely, each of the three leaflets of Intact No.1 and Intact No. 3 have very different extensibilities. Since leaflet extensibility is highly dependent on fixation pressure, it may be that Intact No.2 and Intact No.4 were fixed at slightly different pressures. A large tissue extensibility, however, can be inadvertently measured if the cusp has a wrinkle that has been "fixed" into the material during manufacture. Since the wrinkle can straighten out elastically as the tissue is stretched, the extensibility of the leaflet is overestimated. The wrinkle, however, reappears when the material relaxes and therefore participates in the functional elastic behavior of the leaflet, much like the natural waviness of the individual collagen fibers. It is likely that Intact No.1 and Intact No.3 had such "elastic" wrinkles in some leaflets. Review of procedural notes does indeed reveal that Intact No. 1 was visibly crumpled, whereas Intact No.4 was unusually free of surface details and may have been prestretched during fixation. The extensibility of Intact No.4 was both significantly lower than those of the other Intact valves (p < 0.00 1) and very similar to that of the low-pressure fixed xenograft (14.9% versus 12.4% strain). Since both bending stiffness and extensibility varied considerably in Intact No.1 and Intact No.3, these two parameters were analyzed for degree of correlation. However, no significant correlation between the bending stiffness of the leaflet and its extensibility was detected. Because, as Fig. 5, Band C, indicates, the buckling measurements also had considerable variability, much greater than had those of either the fresh or low-pressure
Analysis of Medtronic Intact valve
97
INTACT 1/ 2 INTACT 1/ 3 INTACT 1/ 4
-
0.6 I
t-
0-
w
0.5
•
Cl
<:>
z
•
o
0.4
::J
:.:: u
0.3
-' « z
0.2
::::J III
0
;:::
u
«
'"
0.1
u,
0.0 0.0
••
•~
0
o._-.-----;----
..-.--a 0
••---6-oeo •
----- •
•
0.5
»>
R = .65
o 1.0
1.5
THICKNESS x CURVATURE CtjR)
Fig. 7. Plot shows the buckling behavior of individual Intact valves and the correlation between fractional buckling depth and thickness X curvature (significant, with p < 0.001). Intact No. 1 was not available for analysis.
fixed tissues, the buckling measurements were coded for the valve to which they correspond and were replotted (Fig. 7). No individual Intact valve demonstrated a buckling behavior different from that of the others. Regression analysis shows that buckling depth correlates with thickness x curvature (p < 0.(01) as expected. The thinner the tissue is or the less it is bent, the lower are the induced strains and the less it buckles. Regression analysis was also applied to buckling measurements to determine whether they were related to the bending stiffness of the material. Although it was expected that the stiffer tissue would buckle more, the extent of buckling appeared unrelated to bending stiffness. It is puzzling to see that, unlike the fresh or low-pressure fixed tissue, the Intact valve demonstrated substantial variability in extensibility, bending stiffness, and buckling behavior. It is therefore likely that some uncontrolled variable affected the measurements. The buckling behavior, in particular, may depend on local fiber densities, and would thus be affected by the presence or absence of major circumferential collagen fiber bundles in the test strip. Discussion The major challenge in developing a new bioprosthetic valve is to increase its durability by reducing calcification and leaflet tearing. Although the long-term performance of the Intact valve can be established only through clinical follow-up, an analysis of its mechanical behavior can offer important insights into its prospective durability. We can rank the three classes of bioprostheses in order of increasing durability as pericardial valve, porcine xenograft, and aortic allograft. These valves have survival rates of 80% at 7 years (pericardialj.P 93% at 7 years (xenograft),22 and 90% and 100% at 10 years (fresh and
98
The Journal of Thoracic and Cardiovascular Surgery
Vesely
cryopreserved aortic allograftj.P The elastic behavior of the most durable of these valves, the allograft, most closely matches that of the natural aortic valve, 14 whereas glutaraldehyde-fixed pericardium least closely duplicates the mechanics of the natural aortic valve. We can therefore deduce that the bioprosthetic valve whose function and tissue characteristics most closely mimic the mechanics of the natural aortic valve willdemonstrate the greatest durability. The analyses of the Intact valve suggest that in general its mechanical behavior is better than that of the lowpressure fixed xenograft. It is, however, not identical to the natural aortic valve. The Intact valve has more "natural" extensibility, stress relaxation, and pliability than has the low-pressure fixed xenograft, yet it buckles the same as the low-pressure fixed xenograft, and more than the natural aortic valve tissue, and it does not relax as quickly as fresh aortic valve tissue. Surprisingly, its elastic modulus is lower than that of the natural aortic valve material. This may be an example of a little-studied but interesting phenomenon of glutaraldehyde fixation relating to tissue shrinkage. A reduction in elastic modulus could result if the leaflets thicken as they shrink during fixation. Although shrinkage during glutaraldehyde fixation is well known, the reasons for its occurrence are not completely understood. Both Lee and coworkers? and Trowbridge and Crofts I6 have noted shrinkage in porcine leaflets and bovine pericardium after fixation. Trowbridge and Crofts 16 reported a decrease in length of 11% after fixation, although neither group reported a change in thickness. If the decrease in length of the tissue produces an increase in the waviness of collagen fibers, then an increase in tissue thickness may occur to accommodate this geometry. Since a given number of fibers must resist the tensile force during stretching, a thickened leaflet therefore has fewer fibers per cross-sectional area than has a thin one. Division of the measured load by the product of specimen width and thickness to obtain stress (0- = l/w . t) therefore yields a proportionately lower stress for a thicker leaflet. The slope of the stress/strain curve, the elastic modulus, is therefore lower. Although leaflet shrinkage during glutaraldehyde fixation is normally associated with increased extensibility.f- 16 it may also be responsible for this decrease in elastic modulus. This, of course, can occur only at fixation pressures low enough to enable significant leaflet contraction. Alternatively, the increase in leaflet thickness after fixation may be only apparent. Fresh tissue is very fluid and may unwittingly be compressed with a thickness-measuring instrument, whereas fixed tissue is more rigid and does not yield when measured. The measured thickness of fixed tissue may therefore be more accurate, and the thickness
of fresh tissue may be consistently underestimated. The net effect is that the measured mean thickness of fixed tissue is greater than that of the fresh. It has been hypothesized that compressive buckling of bioprosthetic tissues leads to delamination, stress concentrations, and eventual tearing of the cuspal material. 11 Since the Intact valve has demonstrated a propensity to buckle during enforced bending, it will likely experience in vivo failure similar to that of the low-pressure fixed xenograft valves. For example, a 0.36 mm thick leaflet bent to a curvature of only 0.67 mm- I (radius of 1.5 mm) buckled to a depth of 0.06 mm. Such curvatures are readily induced in porcine xenograft leaflets during systolic valve opening, as seen in a pulse tank. Whether the reduced elastic modulus, the increased extensibility of wrinkled leaflets, or the overall high variability of the Intact valve has any detrimental effects on long-term function, however, is a matter of speculation. In fact, the Intact valve appears to be a step in the right direction toward the development of a bioprosthetic valve that truly duplicates the natural aortic valve. In spite of high variability, all of the Intact valves tested had elastic characteristics equivalent to or better than those of the low-pressure fixed xenograft. It has therefore been shown that commercial fixation of aortic xenograft valvesat zero pressure preserves many of the beneficial elastic characteristics of porcine aortic valves that ultimately determine their susceptibility to mechanical damage. The author thanks Medtronic of Canada Ltd. for providing the Intact valves used in this study. It should be noted, however, that Medtronic did not contribute any funds toward this research. REFERENCES I. Gallo I, Ruiz B, Nistal F. Degeneration in primary bioprosthetic cardiac valves: incidence of primary tissue failures among 938 bioprostheses at risk. Am J Cardiol 1984;53:1061-5. 2. Milano A, Bortolotti U, Talenti E. Calcific degeneration as the main cause of porcine bioprosthetic valve failure. Am J Cardiol 1984;53: 1066-70. 3. Bortolotti U, Milano A, Mazzucco A, et al. Results of reoperation for primary tissue failure of porcine bioprostheses. J THORAC CARDIOVASC SURG 1985;90:564-9. 4. Schoen FJ. Cardiac valve prostheses: review of clinical status and contemporary biomaterial issues. J Biomed Mater Res 1987;21:91-117. 5. Gabbay S, Kadam P, Factor S, Cheung TK. Do heart valve bioprostheses degenerate for metabolic or mechanical reasons? J THORAC CARDIOVASC SURG 1988;55:208-15. 6. Ishihara T, Ferrans V J, Boyce SW, Jones M, Roberts We. Structure and classification of cuspal tears and perforations
Volume 101
Analysis of Medtronic Intact valve
Number 1 January 1991
7.
8.
9.
10.
11.
12.
13.
14.
in porcine bioprosthetic cardiac valves implanted in patients. Am J Cardiol 1981;48:665-78. Thubrikar MJ, Aouad J, Nolan SP. Patterns of calcific deposits in operatively excised stenotic or purely regurgitant aortic valves and their relation to mechanical stress. Am J Cardiol 1986;54:304-8. Broom NO. The stress/strain and fatigue behaviour of glutaraldehyde preserved heart valve tissue. J Biomech 1977;10:707-24. Lee JM, Boughner OW, Courtman OW. The glutaraldehyde-stabilized porcine aortic valve xenograft. 11. Effect of fixation with or without pressure on the tensile viscoelastic properties of the leaflet material. J Biomed Mater Res 1984;18:79-98. 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. J Biomech 1989;22:655-71. Vesely I, Boughner DR, Song T. Tissue buckling as a mechanism ofbioprosthetic valve failure. Ann Thorac Surg 1988;46:302-8. Broom NO, Thomson F J. Influence of fixation conditions on the performance of glutaraldehyde-treated porcine aortic valves: towards a more scientific basis. Thorax 197934: 166-76. Imamura E, Ishihara S, Ohteki H, Aomi S, Koyanagi H. Open-position fixation of bioprostheses for more physiological performance. J THoRAc CARDIOVASC SURG 1984; 88:114-21. Vesely I, Gonzalez-Lavin L, Graf 0, Boughner DR. Mechanical testing of cryopreserved aortic allografts: com-
15.
16.
17.
18.
19. 20. 21. 22.
23.
99
parison with xenografts and fresh tissue. J THORAC CARDIOVASC SURG 1990;99: 119-23. Lee JM, Courtman OW, Boughner DR. The glutaraldehyde-stabilized porcine aortic valve xenograft. I. Tensile viscoelastic properties of the fresh leaflet material. J Biomed Mater Res 1984;18:61-77. Trowbridge EA, Crofts CEo The standardisation of gauge length: its influence on the relative extensibility of natural and chemically modified pericardium. J Biomech 1986;19: 1023-33. Rousseau EPM, Sauren AAHG, van Hout MC, van Steenhoven AA. Elastic and viscoelastic material behavior of fresh and glutaraldehyde-treated porcine aortic valve tissue. J Biomech 1983;16:339-48. Sauren AAHJ, van Hout MC, van Steenhoven AA, Veldpaus FE, Janssen JD. The mechanical properties of porcine aortic valve leaflets. J Biomech 1983;16:327-37. Vesely I, Boughner DR. A multipurpose tissue bending machine. J Biomech 1985;18:511-3. Zar JH. Biostatistical analysis. Englewood Cliffs, NJ: Prentice-Hall, 1984. Timoshenko SP, Gere JM. Mechanics of materials. New York: Van N astrand Rienhold Co, I972. Nistal F, Garcia-Satul E, Artinano E, Duran CMG, Gallo I. Comparative study of primary tissue valve failure between Ionescu-Shiley pericardial and Hancock porcine valves in the aortic position. Am J Cardiol 1986;57: 161-4. O'Brien MF, Stafford G, Gardner M, et al. A comparison of aortic valve replacement with viable cryopreserved and fresh allograft valves, with a note on chromosomal studies. J THORAC CARDIOVASC SURG 1987;94:812-23.
THORAC CARDIOVASC SURG
1991;101:90-9
Analysis of the Medtronic Intact bioprosthetic valve * Effects of "zero-pressure" fixation The long-term performance of current-design porcine xenograft valves bas not been satisfactory. These valves are generally fixed at "low pressures" of about 3 to 5 mm Hg. The Medtronic Intact (Medtronic, Inc., Minneapolis, Minn.) valve is fixed at "zero pressure" and is proposed as a better alternative to existing xenograft valves. A mechanical analysis of this valve bas been carried out to determine if the Intact valve differs significantly from the low-pressure fixed xenograft Twelve circumferential strips of tissue 5 nun wide were cut from the leaflets of four clinical-grade Intact valves. Their stress/strain, stress relaxation, and flexural bebavior were examined mechanically and histologically. The Intact valve was more extensible tban the low-pressure fixed xenograft (22 % versus 12 % strain, P < 0.001~ relaxed faster (p < 0.001~ and was more pliable tban the xenograft (p < 0.05). It did not, bowever, buckle less tban did the low-pressure fixed xenograft during enforced bending, and it buckled significantly more tban did fresb porcine aortic valve tissue (p < 0.001). The Intact valve also relaxed significantly more slowly tban did the fresb tissue (p < 0.05). Its bending stiffness bad a stronger dependence on leaflet thickness tban the bending thickness of fresb tissue bad (p < 0.001) but a weaker dependence tban the bending thickness of the low-pressure fixed xenograft material bad (p < 0.001). The Intact valve demonstrated a very large variabilitiy in extensibility, bending stiffness, and buckling bebavior, witb little correlation between these parameters. Some valves appeared to bave wrinkled leaflets; others were likely fixed at different pressures. The shrinkage of the leaflet material at these low fixation pressures is likely important, since it can modify the elastic bebavior of the valve cusps. Overall, the Intact valve bad a more ''natural'' elastic bebavior tban bad low-pressure fixed xenograft, and it should therefore experience lower stresses during normal valve function. It can be concluded tbat zero-pressure fixation does preserve many of tbe desirable stress-reducing properties of aortic valve tissue.
Ivan Vesely, PhD, London. Ontario. Canada
Long-term clinical studies!" of bioprosthetic valve implants have demonstrated that the major deficiency of these devices is their poor durability. Although 95% still function after 5 years, as few as 40% remain functional From The John P. Robarts Research Institute, London, Ontario, Canada. Supported by a fellowship from the Heart and Stroke Foundation of Canada. Received for publication July 24, 1989. Accepted for publication Feb. 2, 1990. Address for reprints: Ivan Vesely, PhD, The John P. Robarts Research Institute, P.O. Box SOlS, London, Ontario, Canada, N6A 5K8. *Medtronic, Inc., Minneapolis, Minn.
12/1/20222
90
after 15 years.' The degenerative failure that produces this attrition results from the tearing and calcification of the valve leaflet material.' The exact mechanism of tissue degeneration remains unclear, but there is strong evidence that mechanical factors, particularly flexural fatigue, play an important role both in leaflet tearing and in calcification. 5-7 Previous analyses of porcine aortic valve leaflets have shown that current "low-pressure" glutaraldehyde-fixation techniques decrease tissue extensibility':" and shearing.!" and increase bending rigidity'? and the propensity to collapse and buckle during leaflet bending. I J The lowering of transvalvular pressures during glutaraldehyde fixation has been shown to preserve more of the natural extensibility and pliability of the leaflet materi-
Volume 101 Number 1
Analysis of Medtronic Intact valve 9 I
January 1991
400
300
FRESH PORCINE AORTIC VALVE TISSUE
CRYOPRESERVED AORTIC ALLOGRAFT
0
o, .x: c en
"'
200
QJ ....
en
100
o +---'~~--r--;;;;;";;;~::"'-~~----.-~~-.---l o 40 30 10 20
10
0
20
30
40
20
30
40
400 - r - - - - - - - - - - - - - - - - - - ,
o
LOW PRESSURE FIXED XENOGRAFT
300
o, .x: c
INTACT VALVE
::l 200 ~
en 100
10
20
10
30
Strain in %
Strain in %
Fig. I. Summary of stress/strain curves for eryopreserved aortic allograft tissue, fresh porcine aortic valve tissue, low-pressure fixed porcine xenograft material, and the Intact valve. Measurements for the first three were obtained previously!" and are included for comparison. Note the lower extensibility of the low-pressure fixed xenograft.
a1.12 It has even been suggested that open-position fixation produces a graft that functions better and may therefore last longer.' J Although most valve manufacturers have responded by lowering fixation pressures, the 3 to 5 mm Hg pressure they use to produce well-shaped cusps remains sufficiently high to significantly reduce leaflet extensibility and pliability.!? 14 Medtronic, Inc. (Minneapolis, Minn.) has recently introduced the Intact (Xenotech) porcine bioprosthesis, a valve that claims "zero-pressure" fixation. Fixing valves at zero pressure has been suggested as a better alternative to low-pressure fixation, since much of the collagen crimp can be preserved. I 2 This material should therefore demonstrate much of the pliability and extensibility of natural aortic valve leaflets, experience lower stresses during valve function, and hence last longer. A mechanical analysis of the Intact valve is therefore warranted to determine how well the mechanical behavior of the leaflet material has been preserved with zero-pressure fixation.
Methods Data acquisition. Twelve strips of tissue were obtained from the leaflets of four clinical-grade Intact porcine bioprosthetic valves.Tensile stress/strain and stress relaxation tests were performed with an Instron tensile testing machine (model 1125, Instron Corp., Canton, Mass.). The tissue was repeatedly stretched at 10 rum/min to a maximal load of 0.98 N (100 gm) to precondition the material and produce reproducible load/ elongation curves.l'' The final load/elongation curves were digitized and transformed to stress/strain curves to normalize for variable length and thickness. Both load and tissue elongation were obtained directly from the chart paper during digitizing on a Summagraphics palette (model MM 1812, Fairfield, Conn.). The errors that may have occurred during plotting and digitizing were therefore negligible compared with those potentially introduced during measurement of specimen width, length, and thickness. The precision in cutting the tissue strips to width was estimated to be ±0.2 mm (4% for as mm wide strip). Thickness was measured with a special gauge constructed for this purpose." and gauge length was measured with a calibrated scale affixed to the tissue-clamping mechanism. The precision of the thickness measurements was estimated to be ± 0.05 mm and that of the gauge length to be ± 0.2 mm. The tissue strips were clamped between machined stainless-steel jaws that were lined
92
The Journal of Thoracic and Cardiovascular Surgery
Vesely
100 90 0\ C
'c '0
E
80
Q)
L-
en en
.... Q)
70
L-
ill1
FRESH PORCINE AORTIC VALVE TISSUE
CRYOPRESERVED AORTIC ALLOGRAFT
(f)
60 50 0.1 100
1.0
10.0
100.0
1000.0 0.1
-
1.0
10.0
100.0
1000.0
10.0
100.0
1000.0
90 0\
c
'c '0
E Q)
80
L-
en en
.... Q)
70
L-
(f)
ill1
60 50 0.1
LOW PRESSURE FIXED XENOGRAFT
1.0
10.0
INTACT VALVE
100.0
1000.0 0.1
TIme in seconds
1.0
Time in seconds
Fig. 2. Summary of stress relaxation curves for cryopreserved aortic allograft tissue, fresh porcine aortic valve tissue, low-pressure fixed porcine xenograft material, and the Intact valve. Measurements for the first three were obtained previously!" and are included for comparison. Note the lower relaxation slopes of the two glutaraldehyde- . fixed tissues.
with No. 320 emery cloth to prevent slippage. Admittedly, tensile testing of cut and clamped tissue strips can introduce certain procedural errors. Tensile testing of this type, however, has become an accepted standard, and it is used routinely to assess the mechanical properties of various bioprosthetic tissues" 16-18 It is particularly valid when different tissue types are compared under identical conditions in the same laboratory. For stress relaxation testing, the specimens were held elongated and the.induced tensile force was recorded over a 5-minute interval. Since previous experience with aortic valve tissue has shown that relaxation rates do not vary with applied stress, we did not control for maximum stress level. Each specimen was allowed to relax from a constant load of 0.98 N. All strips were 5 mm wide, cut in the circumferential direction, and tested in a physiological saline bath heated to 37° C. All stress calculations are based on "engineering stress," not "true stress," using the original tissue thickness instead of accounting for thinning during elongation. This approximation avoids estimating the Poisson ratio and the lateral contraction of the specimen during the course of the test. Both of these estimates are required to closely approximate the exact specimen cross-sectional area at any given strain. A more detailed description of the testing procedure can be found in the report by Vesely and colleagues."
Bending tests were performed on the same 12 strips using a custom-built tissue-bending machine. 19 In a heated saline bath, the strips were bent through a 9Q-degreearc to varied curvatures and their bending moments measured to an accuracy of I fJ.N . m (0.1 gm . mm). Each strip was tested four times to establish repeatability." Following bending tests, the strips were cut in halflengthwise, held bent with surgical clips, histologically processed, sectioned to a thickness of 7 fJ.m, stained with van Geison's stain, and examined with low-power light microscopy to measure the degree to which the tissue buckled during enforced bending. I I Data for all tests were compared with data for fresh porcine aortic valve tissue obtained from a local abattoir, low-pressure fixed porcine xenograft tissue obtained from St. Jude/Biolmplant (St. Hyacinthe, Quebec, Canada), and cryopreserved human aortic valve tissue obtained from the Deborah Heart and Lung Center (Browns Mills, N.J.). Data for these tissues had been acquired and published previously." Data analysis. The elastic modulus of each strip was calculated froin the stress/strain curve at a stress of 300 kPa. This stress level simulated the tension the leaflets would experience during diastolic valve closure.l'' The length of the "low-modulus phase" of each stress/strain curve, the leaflet extensibility,
Volume 101 Number 1
Analysis of Medtronic Intact valve
January 1991
93
40,--------------------,
o
N
E
!
CRYOPRESERVED AORTIC ALLOGRAFT
A INTACT VALVE
30
ff) ff)
w
i'u,f.
20
ti (:J
~ 10 z
W
rn O+--t__---+--+--+-+--+-~t____+-_+_-__1
A 0.0
0.1
0.2
0.3
0.4
0.5 0.6 (mm)
0.7
0.8
0.9
1.0
THICKNESS
40,-----------------------,.,-,
40 N
•
E
FRESH PORCINE AORTIC VALVE TISSUE
z 30 --S
A INTACT VALVE
ff) ff)
.. . _-e:-.. --" ---.--
w
zu,
u, ~
20
ff)
•
z 10 z w
rn
a
6.
Z 30 --S
.... INTACT VALVE
E
0.1
0.2
0.3
0.4
0.5 0.6 THICKNESS (mm)
LOW PRESSURE FIXED XENOGRAFT
ff) ff)
w
i'u,f.
T
ff)
(:J
1
0.7
20
~
A
_ .:.---
(:J
0
B 0.0
N
~ 10
z
W
m
0.8
0.9
1.0
c
a +--+--+---f''-----i''''=-t--+----l--+--f----J 0.0
0.1
0.2
0.3
0.4
0.5 0.6 (mm)
0.7
0.8
0.9
1.0
THICKNESS
Fig. 3. Bending stiffness plots of the Intact valve compared with the cryopreserved aortic allograft (A),fresh porcine aortic valve tissue (B), and low-pressure fixed porcine xenograft material (C). The data points and error bars represent the mean and standard deviation of bending-stiffness measurements obtained from four bending curves for each tissue strip. The lines of best fit were obtained by least-squares regression and demonstrate the dependence of bending stiffness on leaflet thickness.
Table I. Calculated values from the curves in Figs. 1 and 2 Elastic modulus at a stress of 300 kPa (MPa)
Summary of tensile data (mean ± SD) I. Allograft 9.05 ± 5.35 2. Fresh tissue 13.01 ± 1.74 3. Xenograft 12.53 ± 3.02 4. Intact valve 9.68 ± 1.64 Comparison of groups* I vs 4 NSt 2 vs 4 p ~ 0.0025 3 vs 4 P ~ 0.05
Length of lowmodulus phase (%) 23.4 22.2 12.4 21.4
± ± ± ±
2.07 3.31 0.87 5.51
NS NS p
~
Slope of stress relaxation curve -5.9 -6.9 -3.5 -4.9
± 2.9
± 1.8 ± 0.8 ± 1.0
NS p
0.0001
~
0.05
p~O.OI
*(iroups "ere statistically compared with a small-sample Student's t test. t:\ot sig nilicaru.
was determined from the intersection of the x axis with a tangent to the stress/strain curve at the 300 kPa point. This parameter indicates the degree to which collagen crimp has been affected by the treatment technique. For example, when porcine valves are fixed at higher pressures, collagen crimp is eliminated and the tissue becomes less extensible. The slopes of the stress relaxation curves were calculated to evaluate integrity of the mucopolysaccharide matrix in each case. A steep slope indi-
cated rapid relaxation of the material through the realignment or migration of fibers in response to applied stress. Such relaxation is possible only if the ground substance is preserved during fixation. Slow relaxation indicates that the material behaves more like an elastic solid, which suggests that the fibers have become immobilized. Measurements of bending moment versus curvature produced bending curves, the slopes of which gave the bending stiffness. Measurements of bending stiffness were first
94
The Journal of Thoracic and Cardiovascular Surgery
Vesely
Fig. 4. Low-power photomicrograph of a typical thin section of buckled Intact valve tissue. Buckling depth was defined as distance (d) as shown, and normalized for the thickness of the specimen (t) to give fractional buckling depth. Buckling depth and tissue thickness were measured with a light microscope and calibrated eyepiece, and the radius of curvature (R) was measured directly from the mounted specimens by means of a hand-held optical comparator with calibrated arcs engraved on the objective. 10 Marker represents 0.1 mm
compared with analysis of variance. The bending stiffness was then plotted against the thickness of the tissue strip, and the slopes of the least-squares regressions of these data were compared by means of multiple Students's t tests. The extent of buckling was assessed by examining the thin sections with lowpower microscopy to measure buckling depth as described previously. I I The radius of curvature, tissue thickness, and buckling depth were measured at several areas around the bend for each of the two pieces cut from the Intact valve test strip. Buckling behavior was expressed as fractional buckling depth versus thickness X curvature, since an increase in both leaflet thickness and in bending curvature yields a corresponding increase in thedegrcc of buckling. I I The lines of best fit through the data sets were compared for elevation as described by Zar. 20 Some errors could have resulted from our methods. The Intact valves were tested for bending stiffness approximately I year after the other material was tested. Although it is conceivable that our bending machine may have produced slightly different results after this period of time, we have not found significant differences between identical fixed tissue tested in 1987 and at present, which makes this an unlikely source of error. Our technique of measuring the gauge length is capable of overestimating tissue extensibility. All specimens, however, were analyzed with the same technique. Our results are therefore internally consistent, and comparisons based on these results are valid. It should be noted, however, that the high variability in material-testing protocols makes the comparison of published studies difficult. 16 Admittedly, measurement of buckling depth
involved some subjectivity. The technique, however, was standardized to minimize this subjective influence, 11 and all tissue was examined by the same individual to control for this influence. Comparisons based on these measurements are therefore valid.
Results Tensile tests. The stress/strain and stress relaxation curves are shown in Figs. 1 and 2 for comparison of the Intact valve, low-pressure fixed xenograft, cryopreserved aortic allograft, and fresh pig aortic valve tissue. From Fig. I, it can be seen immediately that the low-pressure fixed xenograft is less extensible than the other three and that the Intact valve has a rather large spread in the extensibilities of individual samples. For detailed comparison, Table I shows a summary of the calculated parameters obtained from the tensile tests. Cryopreserved aortic allografts and fresh pig aortic valve tissue had similar elastic moduli at high stress, a similarly extensibility (length oflow-modulus phase), and nearly identical rates of stress relaxation. There was no significant difference between these two groups of unfixed tissues. Although the low-pressure fixed porcine xenograft leaflets had the same elastic modulus, they were much less extensible than
Volume 101
Analysis of Medtronic Intact valve
Number 1 January 1991
U.b , - - - - - - - - - - - - -
~
0.5
(:>
0.4
w
c: •
c:
~ ~
-----
0.3
U
OJ OJ ...J
-c z a
0.2
Ci
0.1 {,.{,.
;:: u
"-
A
0.0
---,
LOW PRESSURE FIXED XENOGRAFT FRESH PORCINE AORTIC VALVE TISSUE
o z
95
-@':-~/~:/;~~-
li. ....- . {,. A--
----
.......
-. - ...
•_e_---.
{,.
-.
_ .... _ - - ;
.---::
+--+-4IHt--+-----~>___jf______<~~~__+~-----J
0.5
0.0
1.0
1.5
THICKNESS x CURVATURE
0.6....------------------,
0.6 I f-
Cl.
w
0
(:>
z
0.5
•
FRESH PORCINE AORTIC VALVE TISSUE
•
INTACT VALVE
I f-
Cl.
u
OJ OJ ...J
« z
0.4
(:>
z
u -c a:: u,
0.4
~
u
0.3
OJ OJ
6t _.. . _.. . A
...J
0.2
;:: u
0.1
-c a::
u,
0.0
B 0.0
0.5
1.0
1.5
THICKNESS x CURVATURE
0.0
C
•
6.
"6-.-.. . l::J.
...__ ~_\.o'"
•
- --
--;:------
c:
~ _----
D. A li.
..._ ....::;....~~
{,..
0
0.1
..
. ... ----......---.... • IJ;,.
0.3
-c z
0.2
INTACT VALVE li.
~
0
;::
•
W 0
~ ~
{,. LOW PRESSURE FIXED XENOGRAFT
0.5
a--{:, {,.
• +-.........------..~__+_~-~__+_~~~-_+_-+---l
0.0
0.5
1.0
1.5
THICKNESS x CURVATURE
Fig. 5. Plots of fractional buckling depth (depth/tissue thickness) versus thickness X curvature for low-pressure fixed xenograft material compared with fresh tissue (A),and the Intact valve compared with fresh tissue (B)and with the xenograft (C). For the Intact valve, up to four buckling measurements were obtained from a single leaflet. The fresh and low-pressure fixed valve tissue was tested previously and is included only for comparison. I I
either the fresh porcine or the allograft tissue (12% versus 22%). This shows that even low-pressure glutaraldehyde fixation reduces tissue extensibility. The cross-linkage "freezes" the normally wavy fibers in their straightened configuration, and subsequent loading of the fixed material produces lower extensions. Fig. 2 and comparisons in Table I show that the fixed xenograft valves did not relax as quickly as the other three. This suggests that the mucopolysaccharide matrix has been either partially removed or cross-linked, so that a rapid reconfiguration of fiber geometry in response to stress cannot occur. The Intact valve had a significantly lower elastic modulus than those of the fresh and low-pressure fixed porcine valves and was more extensible than other xenograft tissue (Table I). Its extensibility was very similar to that of the fresh tissue. This would suggest that zero-pressure fixation does indeed preserve the natural extensibility of the aortic leaflet by fixing the collagen fibers in their naturally wavy configuration. Although the Intact valve relaxed more quickly than did the low-pressure fixed xenograft (relaxation slope of -4.9 versus -3.5,
p < 0.01), it relaxed more slowly than did the fresh tissue (slope of -4.9 versus -6.9,p < 0.05). The viscous behavior of the Intact valve is therefore better than that of conventional xenografts but still not as good as that of natural aortic valve tissue. It is interesting to note that the relaxation of the Intact valve was not significantly different from that of the cryopreserved allograft. It is likely that storage and freezing of the allograft valves facilitates the loss of some ground substance, just as does glutaraldehyde fixation. Bending tests. The results of the bending-stiffness measurements are shown in Fig. 3. Analysis of variance and a N ewman-Keuls test show no differences in bending stiffness among the Intact valve, the cryopreserved allograft, and the fresh pig valve tissue. The low-pressure fixed xenograft, however, was significantly stiffer than the others (p < 0.05). Zero-pressure fixation therefore produces a valve leaflet with a pliability comparable to that of unfixed tissue and superior to that of low-pressure fixation. It is interesting to note that thin Intact valve leaflets are more pliable than unfixed tissue but thicker leaflets are stiffer than unfixed tissue. This dependence of
The Journal of
96
Vesely
Thoracic and Cardiovascular Surgery
400.,-------------------,
300
INTACT # 2
o .x:
0.. C
.;;; 200 (J)
~
Vi 100
o +-.--.........~=;:::;:::;::::;=....~~-,---.-~..-.----1 o 40 10 30 20
0
10
20
30
40
400.,------------------,
300 o 0.. .x: c
INTACT # 3
INTACT # 4
';;;200 (J)
~
Vi
100
10
20 Strain in %
30
10
20 Strain in %
30
40
Fig. 6. Stress/strain curves for the four Intact valves tested. Note the presence of both intervalve and intravalve variability of leaflet extensibilities. Although the leaflets of Intact No.2 and Intact No.4 have similar extensibilities, . the two valves appear to have been fixed at different pressures. The high variability within Intact No. I and Intact No.3 probably results from wrinkled leaflets.
bending stiffness on leaflet thickness was analyzed by obtaining least-squares regressions for the four groups of bending-stiffness data and comparing them with multiple t tests. The Intact valve had a significantly greater slope (p < 0.(01) than had both the cryopreserved allograft (Fig. 3, A) and the fresh porcine aortic valve (Fig. 3, B). This slope, however, was lower (p < 0.001) than that for the low-pressure fixed xenograft (Fig. 3, C). Fixed tissue therefore demonstrated a greater increase of bending stiffness for a given change in thickness than did unfixed tissue, and this relationship intensified with fixation pressure. It should be noted that there is only a 0.006 probability of type I error with these multiple t tests. 20 Bending stiffness also depends both on the degree of cross-linkage between fiber layers and on the extensibility of the tissue. Cross-linkage increases bending stiffness by preventing shearing between fiber layers.!? and a reduced extensibility means that the material operates at a high modulus, even at the low strains occurring during
bending. A reduction in ability to shear also makes the material behave more like an elastic solid, with a bending stiffness strongly dependent on the sectional modulus (the second moment of area)." In such an "ideally elastic" material, the bending stiffness increases with the cube of the thickness, yielding a very high dependence of bending stiffness on leaflet thickness.'? The low glutaraldehyde concentrations used to fix the Intact valve may well have better preserved the ability of the valve to shear during bending, hence increasing its pliability. Buckling tests. A typical thin section of buckled Intact valve tissue is shown in Fig. 4. The analyses of such histologically prepared sections are shown in Fig. 5. Fig. 5, A, is a comparison plot of the buckling behavior of lowpressure fixed xenograft tissue and fresh porcine aortic valve tissue (obtained previously). II The difference in elevation of the two regression lines demonstrates that the xenograft buckled more than did the fresh tissue at all curvatures. This difference was significant (p < 0.(01).
Volume 101 Number 1 January 1991
Fig. 5, Band C, compares the Intact valve with the fresh tissue and with the low-pressure fixed xenograft. The Intact valve buckled significantly more (p < 0.001) than did the fresh tissue but not significantly less than did the low-pressure fixed xenograft. Like the xenograft, the Intact valve buckled at curvatures too low to produce any buckling in the fresh tissue. Compressive buckling induced in the leaflet during systolic valve opening has been advanced as a probable mechanism for bioprosthetic tissue disruption and cuspal tearing in vivo.11 The similar buckling behaviors of the Intact and the low-pressure fixed xenograft suggest that implanted Intact valves may also fail in this manner. Variability of measured parameters. The stress/ strain curves in Fig. 1 show that the extensibility of the Intact valve is considerably more variable than that of the others. To examine this variability further, the stress/ strain curves for each valve were plotted separately (Fig. 6). These plots show both intervalve and intravalve variability in the extensibilities ofthe cuspal tissue. Intact No. 2 and Intact No.4 are internally consistent but have different overall extensibilities (p < 0.(025). Conversely, each of the three leaflets of Intact No.1 and Intact No. 3 have very different extensibilities. Since leaflet extensibility is highly dependent on fixation pressure, it may be that Intact No.2 and Intact No.4 were fixed at slightly different pressures. A large tissue extensibility, however, can be inadvertently measured if the cusp has a wrinkle that has been "fixed" into the material during manufacture. Since the wrinkle can straighten out elastically as the tissue is stretched, the extensibility of the leaflet is overestimated. The wrinkle, however, reappears when the material relaxes and therefore participates in the functional elastic behavior of the leaflet, much like the natural waviness of the individual collagen fibers. It is likely that Intact No.1 and Intact No.3 had such "elastic" wrinkles in some leaflets. Review of procedural notes does indeed reveal that Intact No. 1 was visibly crumpled, whereas Intact No.4 was unusually free of surface details and may have been prestretched during fixation. The extensibility of Intact No.4 was both significantly lower than those of the other Intact valves (p < 0.00 1) and very similar to that of the low-pressure fixed xenograft (14.9% versus 12.4% strain). Since both bending stiffness and extensibility varied considerably in Intact No.1 and Intact No.3, these two parameters were analyzed for degree of correlation. However, no significant correlation between the bending stiffness of the leaflet and its extensibility was detected. Because, as Fig. 5, Band C, indicates, the buckling measurements also had considerable variability, much greater than had those of either the fresh or low-pressure
Analysis of Medtronic Intact valve
97
INTACT 1/ 2 INTACT 1/ 3 INTACT 1/ 4
-
0.6 I
t-
0-
w
0.5
•
Cl
<:>
z
•
o
0.4
::J
:.:: u
0.3
-' « z
0.2
::::J III
0
;:::
u
«
'"
0.1
u,
0.0 0.0
••
•~
0
o._-.-----;----
..-.--a 0
••---6-oeo •
----- •
•
0.5
»>
R = .65
o 1.0
1.5
THICKNESS x CURVATURE CtjR)
Fig. 7. Plot shows the buckling behavior of individual Intact valves and the correlation between fractional buckling depth and thickness X curvature (significant, with p < 0.001). Intact No. 1 was not available for analysis.
fixed tissues, the buckling measurements were coded for the valve to which they correspond and were replotted (Fig. 7). No individual Intact valve demonstrated a buckling behavior different from that of the others. Regression analysis shows that buckling depth correlates with thickness x curvature (p < 0.(01) as expected. The thinner the tissue is or the less it is bent, the lower are the induced strains and the less it buckles. Regression analysis was also applied to buckling measurements to determine whether they were related to the bending stiffness of the material. Although it was expected that the stiffer tissue would buckle more, the extent of buckling appeared unrelated to bending stiffness. It is puzzling to see that, unlike the fresh or low-pressure fixed tissue, the Intact valve demonstrated substantial variability in extensibility, bending stiffness, and buckling behavior. It is therefore likely that some uncontrolled variable affected the measurements. The buckling behavior, in particular, may depend on local fiber densities, and would thus be affected by the presence or absence of major circumferential collagen fiber bundles in the test strip. Discussion The major challenge in developing a new bioprosthetic valve is to increase its durability by reducing calcification and leaflet tearing. Although the long-term performance of the Intact valve can be established only through clinical follow-up, an analysis of its mechanical behavior can offer important insights into its prospective durability. We can rank the three classes of bioprostheses in order of increasing durability as pericardial valve, porcine xenograft, and aortic allograft. These valves have survival rates of 80% at 7 years (pericardialj.P 93% at 7 years (xenograft),22 and 90% and 100% at 10 years (fresh and
98
The Journal of Thoracic and Cardiovascular Surgery
Vesely
cryopreserved aortic allograftj.P The elastic behavior of the most durable of these valves, the allograft, most closely matches that of the natural aortic valve, 14 whereas glutaraldehyde-fixed pericardium least closely duplicates the mechanics of the natural aortic valve. We can therefore deduce that the bioprosthetic valve whose function and tissue characteristics most closely mimic the mechanics of the natural aortic valve willdemonstrate the greatest durability. The analyses of the Intact valve suggest that in general its mechanical behavior is better than that of the lowpressure fixed xenograft. It is, however, not identical to the natural aortic valve. The Intact valve has more "natural" extensibility, stress relaxation, and pliability than has the low-pressure fixed xenograft, yet it buckles the same as the low-pressure fixed xenograft, and more than the natural aortic valve tissue, and it does not relax as quickly as fresh aortic valve tissue. Surprisingly, its elastic modulus is lower than that of the natural aortic valve material. This may be an example of a little-studied but interesting phenomenon of glutaraldehyde fixation relating to tissue shrinkage. A reduction in elastic modulus could result if the leaflets thicken as they shrink during fixation. Although shrinkage during glutaraldehyde fixation is well known, the reasons for its occurrence are not completely understood. Both Lee and coworkers? and Trowbridge and Crofts I6 have noted shrinkage in porcine leaflets and bovine pericardium after fixation. Trowbridge and Crofts 16 reported a decrease in length of 11% after fixation, although neither group reported a change in thickness. If the decrease in length of the tissue produces an increase in the waviness of collagen fibers, then an increase in tissue thickness may occur to accommodate this geometry. Since a given number of fibers must resist the tensile force during stretching, a thickened leaflet therefore has fewer fibers per cross-sectional area than has a thin one. Division of the measured load by the product of specimen width and thickness to obtain stress (0- = l/w . t) therefore yields a proportionately lower stress for a thicker leaflet. The slope of the stress/strain curve, the elastic modulus, is therefore lower. Although leaflet shrinkage during glutaraldehyde fixation is normally associated with increased extensibility.f- 16 it may also be responsible for this decrease in elastic modulus. This, of course, can occur only at fixation pressures low enough to enable significant leaflet contraction. Alternatively, the increase in leaflet thickness after fixation may be only apparent. Fresh tissue is very fluid and may unwittingly be compressed with a thickness-measuring instrument, whereas fixed tissue is more rigid and does not yield when measured. The measured thickness of fixed tissue may therefore be more accurate, and the thickness
of fresh tissue may be consistently underestimated. The net effect is that the measured mean thickness of fixed tissue is greater than that of the fresh. It has been hypothesized that compressive buckling of bioprosthetic tissues leads to delamination, stress concentrations, and eventual tearing of the cuspal material. 11 Since the Intact valve has demonstrated a propensity to buckle during enforced bending, it will likely experience in vivo failure similar to that of the low-pressure fixed xenograft valves. For example, a 0.36 mm thick leaflet bent to a curvature of only 0.67 mm- I (radius of 1.5 mm) buckled to a depth of 0.06 mm. Such curvatures are readily induced in porcine xenograft leaflets during systolic valve opening, as seen in a pulse tank. Whether the reduced elastic modulus, the increased extensibility of wrinkled leaflets, or the overall high variability of the Intact valve has any detrimental effects on long-term function, however, is a matter of speculation. In fact, the Intact valve appears to be a step in the right direction toward the development of a bioprosthetic valve that truly duplicates the natural aortic valve. In spite of high variability, all of the Intact valves tested had elastic characteristics equivalent to or better than those of the low-pressure fixed xenograft. It has therefore been shown that commercial fixation of aortic xenograft valvesat zero pressure preserves many of the beneficial elastic characteristics of porcine aortic valves that ultimately determine their susceptibility to mechanical damage. The author thanks Medtronic of Canada Ltd. for providing the Intact valves used in this study. It should be noted, however, that Medtronic did not contribute any funds toward this research. REFERENCES I. Gallo I, Ruiz B, Nistal F. Degeneration in primary bioprosthetic cardiac valves: incidence of primary tissue failures among 938 bioprostheses at risk. Am J Cardiol 1984;53:1061-5. 2. Milano A, Bortolotti U, Talenti E. Calcific degeneration as the main cause of porcine bioprosthetic valve failure. Am J Cardiol 1984;53: 1066-70. 3. Bortolotti U, Milano A, Mazzucco A, et al. Results of reoperation for primary tissue failure of porcine bioprostheses. J THORAC CARDIOVASC SURG 1985;90:564-9. 4. Schoen FJ. Cardiac valve prostheses: review of clinical status and contemporary biomaterial issues. J Biomed Mater Res 1987;21:91-117. 5. Gabbay S, Kadam P, Factor S, Cheung TK. Do heart valve bioprostheses degenerate for metabolic or mechanical reasons? J THORAC CARDIOVASC SURG 1988;55:208-15. 6. Ishihara T, Ferrans V J, Boyce SW, Jones M, Roberts We. Structure and classification of cuspal tears and perforations
Volume 101
Analysis of Medtronic Intact valve
Number 1 January 1991
7.
8.
9.
10.
11.
12.
13.
14.
in porcine bioprosthetic cardiac valves implanted in patients. Am J Cardiol 1981;48:665-78. Thubrikar MJ, Aouad J, Nolan SP. Patterns of calcific deposits in operatively excised stenotic or purely regurgitant aortic valves and their relation to mechanical stress. Am J Cardiol 1986;54:304-8. Broom NO. The stress/strain and fatigue behaviour of glutaraldehyde preserved heart valve tissue. J Biomech 1977;10:707-24. Lee JM, Boughner OW, Courtman OW. The glutaraldehyde-stabilized porcine aortic valve xenograft. 11. Effect of fixation with or without pressure on the tensile viscoelastic properties of the leaflet material. J Biomed Mater Res 1984;18:79-98. 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. J Biomech 1989;22:655-71. Vesely I, Boughner DR, Song T. Tissue buckling as a mechanism ofbioprosthetic valve failure. Ann Thorac Surg 1988;46:302-8. Broom NO, Thomson F J. Influence of fixation conditions on the performance of glutaraldehyde-treated porcine aortic valves: towards a more scientific basis. Thorax 197934: 166-76. Imamura E, Ishihara S, Ohteki H, Aomi S, Koyanagi H. Open-position fixation of bioprostheses for more physiological performance. J THoRAc CARDIOVASC SURG 1984; 88:114-21. Vesely I, Gonzalez-Lavin L, Graf 0, Boughner DR. Mechanical testing of cryopreserved aortic allografts: com-
15.
16.
17.
18.
19. 20. 21. 22.
23.
99
parison with xenografts and fresh tissue. J THORAC CARDIOVASC SURG 1990;99: 119-23. Lee JM, Courtman OW, Boughner DR. The glutaraldehyde-stabilized porcine aortic valve xenograft. I. Tensile viscoelastic properties of the fresh leaflet material. J Biomed Mater Res 1984;18:61-77. Trowbridge EA, Crofts CEo The standardisation of gauge length: its influence on the relative extensibility of natural and chemically modified pericardium. J Biomech 1986;19: 1023-33. Rousseau EPM, Sauren AAHG, van Hout MC, van Steenhoven AA. Elastic and viscoelastic material behavior of fresh and glutaraldehyde-treated porcine aortic valve tissue. J Biomech 1983;16:339-48. Sauren AAHJ, van Hout MC, van Steenhoven AA, Veldpaus FE, Janssen JD. The mechanical properties of porcine aortic valve leaflets. J Biomech 1983;16:327-37. Vesely I, Boughner DR. A multipurpose tissue bending machine. J Biomech 1985;18:511-3. Zar JH. Biostatistical analysis. Englewood Cliffs, NJ: Prentice-Hall, 1984. Timoshenko SP, Gere JM. Mechanics of materials. New York: Van N astrand Rienhold Co, I972. Nistal F, Garcia-Satul E, Artinano E, Duran CMG, Gallo I. Comparative study of primary tissue valve failure between Ionescu-Shiley pericardial and Hancock porcine valves in the aortic position. Am J Cardiol 1986;57: 161-4. O'Brien MF, Stafford G, Gardner M, et al. A comparison of aortic valve replacement with viable cryopreserved and fresh allograft valves, with a note on chromosomal studies. J THORAC CARDIOVASC SURG 1987;94:812-23.