- Email: [email protected]
Pulsatile flow past St. Jude Medical bileaflet valve
J THORAC
CARDIOVASC SURG
89:743-749, 1985
Pulsatile flow past 81. Jude Medical bileaflet valve An in vitro study An in vitro hemodynamic study of the St. Jude Medical bileaflet aortic prosthesis was performed in a mock circulatory system simulating physiological pulsatile flow. The study included measurements of pressure drop across the valves, percent regurgitation, velocity, and turbulence in a model human aorta. The measurements indicated that pressure drop (mean systolic pressure drop of 6.2 mm Hg), percent regurgitation (10.15%), and turbulent normal stresses immediately downstream from the valve (825 dynes/cm~ were better than those with other prosthetic valvesand bioprostheses. The flow development in the aorta was not significantly affected by the orientation of the bileaflet valve in the root of the aorta. However, velocity measurements immediately downstream from the valves showed flow reversal and separation in the vicinity of the binge points of the leaflets where thrombus formation bas been previously reported.
Krishnan B. Chandran, D.Sc., Iowa City, Iowa
h e centrally occluding caged ball valve was introduced as a replacement for diseased aortic valves in 1960. Within a decade, with the introduction ofpyrolytic carbon discs, tilting disc valves gained popularity. More recently, a bileaflet valve geometry has been introduced into the market. The 81. Jude Medical bileaflet valve has two leaflets mounted in an orifice ring all coated with pyrolytic carbon. The leaflets open to about 85 degrees, and there are two major flow orifices and a minor central orifice. The valve has a low profile without any protecting guiding struts on the aortic side. In the closed position, the leaflets do not overlap and leave a ring slit between the leaflets and the orifice ring as well as a central slit where the two orifices meet.' In vitro hemodynamic evaluation of the relatively new geometry for valve prostheses has been previously reported.t' Gombrich, Villafana, and Palmquist/ compared the pressure gradients across the valve during steady and pulsatile flow, the flow dynamics in pulsatile From the Department of Biomedical Engineering, University of Iowa, Iowa City, Iowa Supported in part by Grant HL-26269 from the U.S. Public Health Service. Received for publication Jan. 30, 1984. Accepted for publication May 31, 1984. Address for reprints: K.B. Chandran, D.Sc., 1218 EB, Department of Biomedical Engineering, College of Engineering, The University of Iowa, Iowa City, Iowa 52242.
flow through qualitative flow visualization studies, and the calculated valve orifice areas between the bileaflet valve and other prosthetic and tissue valve. Their results suggested that the bileaflet valve had relatively lower pressure gradients than did other valve geometries, and computed valve orifice areas were larger for the bileaflet valves with comparable tissue anulus diameters. Bruss and associates! compared the pressure drop across the valves during steady flow for bileaflet and tilting disc valves and compared the velocity profiles downstream from the valves under pulsatile flow conditions with laser Doppler anemometry (LOA). However, they did not attempt to measure the fluctuating component of the velocity in their studies. Their results showed that the bileaflet valves had the lowest pressure gradient across the valve under steady flow conditions. These results have been further confirmed by Emery and Nicoloff" and Gabbay and associates.' The velocity profiles during pulsatile flow conditions by Bruss and co-workers! showed that the flow downstream to the bileaflet valve was relatively flat compared to the tilting disc valves and the wall shear stress could also be expected to be the least with the bileaflet valves. More recently, Yoganathan's group':" has performed in vitro pressure-drop studies and velocity measurements downstream from the St. Jude Medical valve under steady flow conditions. Gray and colleagues' have also discussed the in vivo hydrodynamic characteristics of bileaflet, tilting disc, and porcine valves. These results suggest that the 743
The Journal of
744
Chandran
Thoracic and Cardiovascular Surgery
Fig. 1. Orientation of the bileaflet valve with respect to the aortic root. A, Tilt axis in the plane of primary curvature. B, Tilt axis perpendicular to the plane of primary curvature. LCA, Left coronary artery. RCA, Right coronary artery.
bileaflet valve has better flow characteristics than the other valves. Despite the improved hemodynamic characteristics exhibited by the bileaflet valve in in vitro evaluation, thromboembolic complications have been reported in the clinical use of this prosthesis. 8,9 In this paper, a hemodynamic evaluation of the St. Jude Medical bileaflet prosthesis in vitro is reported. The measurements included ·pressure-drop studies, qualitive visualization of flow development past the prosthesis in a model human aorta, and quantitative measurement of velocity profiles and turbulent stresses performed with LDA. The results of the experimental study are compared with those obtained with other types of prosthetic valves as well as bioprostheses. Experimental procedure A St. Jude Medical bileaflet pyrolytic carbon valve with a tissue anulus diameter of 27 mm was used in this study. The valve was sutured to an acrylic valve mounting ring and incorporated into a mock circulatory system simulating physiological pulsatile flow. The mock circulatory system used in our laboratory has been described previously'? and was run at 72 beats/min with a systolic duration of 270 msec. The blood analog fluid used in these experiments was a glycerol solution (36% by volume in distilled water) whose viscosity coefficient and density at room temperature were representative of that of human blood at high rates of shear. Qualitative flow visualization studies were performed with illuminated, neutrally buoyant resin beads in a model human aorta as described by Chandran and associates. I I These studies were followed by quantitative measurements
Fig. 2. Flow visualization in the ascending aorta in early systole with the bileaflet valve. A, Valve oriented with the tilt axis in the plane of primary curvature (orientation A). B, Valve oriented with the tilt axis perpendicular to the plane of primary curvature (orientation B).
with LDA in a forward scattering mode at various cross sections in the ascending aorta and in the mid-arch region. The details of the measurements with LDA in our laboratory have been described elsewhere. 12 For these studies, the bileaflet valve was mounted in two different orientations with respect to the root of the aorta, so that the effect of orientation on the flow development could be analyzed. In orientation A, the tilt axis was in the plane of the primary curvature of the aorta and in orientation B, the tilt axis was perpendicular to the plane of primary curvature, as shown in Fig. I. The pressure-drop measurements across the bileaflet
Volume 89
St. Jude Medical valve 745
Number 5 May. 1985
valve were obtained under physiological pulsatile flow conditions with a Validyne P35d differential pressure transducer. For the pressure-drop studies, an axisymmetric valvular chamber" was incorporated into the mock circulatory system replacing the model human aorta. The axisymmetric flow chamber had pressure taps both upstream and downstream from the valve for the pressure-drop measurements. This chamber was also used to obtain velocity measurements 21 mm downstream from the seat of the valve, whereas in the aorta, data could be obtained no closer to the valve than 28 mm in the experimental setup used in these studies. The measurement 21 mm downstream from the seat of the valve is approximately 13 mm downstream from the valve, whereas Yoganathan and associates'" made the measurements 6 mm downstream from the valve under steady flow conditions. Results Flow visualization study. Flow visualization study was performed by recording the motion of the illuminated resin beads in the aortic flow chamber. Highspeed movies at 100 frames/sec were obtained and the motion of the particles was analyzed with a motion analyzer projector. Also, 35 mm photographs were obtained with a 35 mm SLR camera driven by a motor driveat 3 frames/sec. The motor drive was triggered by the electronic pulser so that the picture could be timed in reference to the cardiac cycle. Fig. 2, A and B shows typical flow visualization pictures in early systole for orientations A and B, respectively. In Fig. 2, A, the valve is oriented such that the minor central flow orifice is in the plane of the picture. In early systole, a uniform forward motion is observed in the ascending aorta. In late systole, this uniform flow is strongly affected by the secondary motion because of a large amount of flow through the major flow orifices near the lateral walls. In early diastole, a flow reversal is initiated along the inner wall of curvature of the aorta (posterior wall), which is soon extended across the cross section in the mid-arch region. In orientation B, with the central flow orifice perpendicular to the plane of the picture, relatively fast fluid motion is observed in the region of major flow orifices, with very little motion in the core region downstream from the central minor orifice. However, as the flow develops, a more uniform forward motion is observed in the ascending aorta. In early diastole, a flow reversal is observed along the inner wall of curvature. A slight forward motion is observed along the outer wall of curvature through a large part of diastole, with a resulting clockwise vortex motion in the plane of the picture in the aortic chamber. In contrast to the study on
20.0 r - - - - - - - - - - - - - - - - ,
[
"'"<
~
3 LL
U 0::
I-
w
15.0
10.0
5.0
:2
::J -.J
'~
0.0 1-------t---f"r----::::;iII"'c---.,
- 5.0
L..-_ _---l.
0.00
0.21
---lIoL-_ _.....L.
0.42 TIME (sec)
0.63
....l
0.84
Fig. 3. A typical instantaneous flow rate signal in a cardiac cycle.
the effect of valve orientation of tilting disc valves,II in neither orientation did the flow development past bileaflet valves appear to produce regions of relative stasis in the ascending aorta or in -the mid-arch region. Pressure-drop and velocity profiles in the axisymmetric chamber. A typical instantaneous flow rate profile in a cardiac cycle as recorded from a Carolina electromagnetic flowmeter is shown in Fig. 3. The cardiac output and percent regurgitation (ratio of retrograde flow over the net positive flow) can be determined by computing the area under the curve for positive and negative flow. In these experiments, the cardiac output was maintained to be about 6 L/m and the percent regurgitation was computed to be 10.15%. This value compared favorably with those for other valve prostheses evaluated in our laboratory. The percent regurgitation was in the range of 10% to 13% for tilting disc valves (Bjork-Shiley and Hall-Kaster), 3% to 10% for caged ball valves (Starr-Edwards), and 3% to 7% for bioprostheses (porcine and pericardial). The peak and the mean systolic pressure drop across the valve were determined with the differential pressure transducer. A mean systolic pressure drop of 6.2 mm Hg was measured for the S1. Jude Medical valve, and the value ranged between 6 and 12 mm Hg for the prosthetic valves. Our results showing one of the lowest pressure drops with the bileaflet valves agreed with those studied reported earlier. I. 2 In the axisymmetric aortic chamber, the velocity profiles 21 mm downstream from the seat of the bileaflet valve were obtained for two different orienta-
The Journalof Thoracic and Cardiovascular Surgery
7 4 6 Chandran
125.0 , - - - - - - - - - - - - - - - - - - - - ,
125.0 , - - - - - - - - - - - - - - - - - - ,
100.0
100.0
E 75.0 VI
~ 75.0 E (J
(J
~ 50.0
50.0
~
U
u
o
o
....J
....J
w 25.0
~ 25.0
>
....J
....J
X
X 0.0
0.0
-25.0 .... --.rL~ -50.0 '--
-25.0 J-
-'--
---'-
-1.0 -0.5 0.0 0.5 NEAR WALL NON-DIMENSIONAL RADIUS
---'
1.0 FAR WALL
Fig. 4. Velocity profiles 21 mm downstream from the St. Jude Medical valve perpendicular to the tilt axis. D = 40 msec after valve opening; V = 160 msec after valve opening; o = 210 rnsec after valve opening; • = 400 rnsec after valve opening.
tions. In the first orientation, the velocity measurements were made by transversing the measuring volume of the LDA perpendicular to the tilt axis of the leaflets. Hence, the velocity measurements were obtained in the two major flow orifices as well as in the central minor flow orifice, In the second orientation, the LDA traverse was parallel to the tilt axis such that the velocity profile in the central flow orifice was obtained. Fig. 4 shows the velocity profiles at different times in the pulsatile flow cycle for the first orientation. In early systole (40 msec after valve opening), a forward flow throughout the cross section is observed, and by peak systole (160 msec), a three-peaked velocity profile in the core region is observed with region of flow reversal and vortex formation in the valvular sinus region. The three-peaked velocity profile corresponding to the three flow regions of the bileaflet valve is very similar to that reported by Bruss et al.' In early diastole, the vortices in the sinus region are still observed while the flow in the core region is reduced and the fluid motion dies out toward the rest of the cardiac cycle in the pulse duplicator. Fig. 5 compares the velocity profile for the two orientations of the valve in peak systole (160 msec). In the orientation parallel to the tilt axis, when the velocity measurement is in the central flow orifice, a single-
-50.0 '--
...l..-
-L
----'-
---'
-1.0 - 0.5 0.0 0.5 1.0 NEAR WALL FAR WALL NON-DIMENSIONAL RADIUS
Fig. 5. Velocity profiles 21 mm downstream from the St. Jude Medical valve in peak systole (160 msec). 0 = Perpendicular to the tilt axis; D = parallel to the tilt axis. peaked velocity profile is observed in the core region with relative stasis in the periphery near the leaflet housing. This region of very little fluid motion is observed for a large portion of the cardiac cycle. Should depositions accumulate in this region, the leaflet motion may be obstructed. Fig. 6 compares the velocity profiles in peak systolic in the axisymmetric chamber for the St. Jude Medical valve with those of Bjork-Shiley convexo-concave tilting disc valve (Model 27 MBRM), a Carpentier-Edwards porcine valve (Model 2625), and an Ionescu-Shiley low-profile pericardial valve. The tilting disc valve with the minor and major flow regions yields a jet-like flow in the major flow orifice with relatively large wall shear stresses and a vortex formation in the minor flow region." With the porcine valve, which has a relatively smaller valve orifice area than the pericardial valve, a strong central core jet is observed with a large vortex formation between the commissure and the valve support, which protrudes into the lumen. The velocity profile for the pericardial valve shows a uniform core flow and a vortex formation between the leaflets and the sinus. The tilting disc valve, along with the caged ball valve, yields flow with relatively large wall shear stresses. The flow past the porcine valve shows a strong central jet with high turbulent stresses." Our measurements showed that the turbulent normal stresses downstream to the valves are about 1,800 to 2,600 dynes/em' for the
Volume 89 Number 5
St. Jude Medical valve
747
May, 1985
porcine valve, about 1,200 to 1,600 dynes/em' for the caged ball, tilting disc, and the pericardial valves, and about 825 dynes/em' for the bileaflet valve. Measurement of turbulent shear stress, which has been directly linked to the damage to formed elements in blood,14,15 requires simultaneous measurement of the axial and radial components of velocity. With a single-channel laser system in our laboratory, it was not possible to measure the turbulent shear stress. Even though the St. Jude Medical valve has one of the lowest pressure gradients of all valve prostheses, along with the relatively smaller turbulent stresses, thromboembolic complications are still reported with this valve.8,9 Velocity profiles in the aorta. Velocity profiles were measured in several cross sections in the ascending aorta, mid-arch region, and in the brachiocephalic artery as shown schematically in Fig. 7. The LDA traverses were made from the right lateral to the left lateral wall in plane AA (28 mm from the seat of the valve), plane BB (48 mm from the seat of the valve), plane CC, and plane DD. Measurements were obtained with the valves mounted in both the orientations described in the flow visualization studies (Fig. 1). Figs. 8 and 9 show the velocity profiles in peak systole (160 msec after valve opening) at cross sections AA and BB, respectively, for both orientations A and B. Even in section AA, 28 mm downstream from the seat of the valve, the velocity profiles are relatively flat. The flow reversal observed near the sinus region or near the leaflet hinges described earlier (Fig. 5) are not apparent in this cross section. In section BB, both the velocity profiles show a tendency to skew toward the left lateral wall, which is the inner wall of the tertiary curvature shown in Fig. 7. However, the flow development in the ascending aorta or in the mid-arch region was not observed to change dramatically with the change in orientation of the bileaflet valve. In the mid-arch region also, the velocity profiles were observed to be relatively flat for both the orientations; in diastole, a flow reversal was observed throughout the mid-arch region with the bileaflet valves. Relatively flat velocity profiles were observed in the brachiocephalic arterial branch (cross section DD) for both orientations of the valve. Discussion The in vitro measurements of pulsatile flow past the St. Jude Medical bileaflet valve suggests that the flow characteristics are comparable, if not superior, to those of prosthetic valves of caged ball and tilting disc geometry. The mean systolic pressure gradient and the percent regurgitation with the bileaflet valves compared favorably with other prosthetic valves as well as biopros-
150,0
120,0
'"
<,
E
90,0
u
>-
t<.)
0
60,0
--.J W
>
--.J
« x «
30.0
0.0
-30,0 ' - - - - _ - - ' - -1.0 - 0, 5 NEAR WALL
---'-0.0
-----'-0.5
-.1
1.0 FAR WALL
NON-DIMENSIONAL RADIUS
Fig. 6. Velocity profiles 21 mm downstream to the valve seat in peak systole. 0 = St. Jude Medical valve; • = Bjork-Shiley convexo-concave valve; 'V = Carpentier-Edwards procine valve; 0 = Ionescu-Shiley low-profile valve,
theses. The measured velocity profiles with the bileaflet valves showed a centralized flow with relatively lower wall shear stresses than with the caged ball and tilting disc valves. The measured turbulent axial stresses were also relatively lower with the bileaflet values than with the other valve prostheses. Flow visualization studies complemented by quantitative velocity measurements in the model human aorta suggested that the orientation of the bileaflet valve at the root of the aorta did not affect the flow development in the ascending aorta. In contrast, with the tilting disc valves, the flow development strongly depended on the orientation of the major flow orifice. I I A comparison of the velocity profiles under pulsatile flow reported in this work with those reported under steady flow" show several interesting features, As pointed out earlier, pulsatile flow measurements were obtained in this study 13 mm downstream from the valve at a peak instantaneous flow rate of 18 Lzrn, whereas the steady flow measurements" were obtained 6 mm downstream from the valve at a flow rate of 25 L/min. The velocity profiles are similar in shape in both orientations for steady and pulsatile flow. However, under pulsatile flow the flow reversal is stronger in the valvular sinus area, with a negative velocity of about 25 em/sec compared to 8 em/sec under steady flow, even though the pulatile flow measurements were made
The Journal of
7 4 8 Chandran
Thoracic and Cardiovascular Surgery
OUTER WALL OF PRIMARY CURVATURE (GREATER CURVE) OUTER WALL OF SECONDARY CURVATURE
+---- DIRECTION OF THE
VELOCITY MEASUREMENT
DIRECTION OF FLOW
Fig. 7. Schema of the model human aorta and the cross sections in which velocity measurements were obtained. 125.0
r------------------....,
100.0 , - - - - - - - - - - - - - - - - - - -
100.0
e'" o
~ 75.0 E
7 5 .O
u
>-
~ 50.0
fooU
o
:l 50.0
~ 25.0
>
-l
-l
u
-l
sx
UJ
0.0
X
-25.0
-50.0 ' - - - - - - - ' - - - - - - - ' ' - - - - - - ' - - - - - - - ' -1.0 -0.5 0.0 0.5 1.0 RIGHT LEFT LATERAL WALL LATERAL WALL
NON-DIMENSIONAL RADIUS
Fig. 8. Velocity profiles in peak systole (160 msec after valve opening) in cross section AA in the ascending aorta. 0 = Orientation A; 0 = orientation B.
0.0 O - - - - - - ' - - - - - - - - ' - - - - - ' - - - - - - a -1.0 - 0.5 0.0 0.5 1.0 RIGHT LEFT LATERAL WALL LATERAL WALL
NON-DIMENSIONAL RADIUS
Fig. 9. Velocity profiles in peak systole (120 msec after valve opening) in cross section BB in the ascending aorta. 0 = Orientation A; 0 = orientation B.
Volume 89 Number 5 May, 1985
further downstream from the valve. The peaks and valleys in the velocity profile in the core region corresponding to the three flow orifices are more prominent in steady flow closer to the valve. Yoganathan and associates" reported maximum turbulent shear stresses of 750 dynes/em- under steady flow. Assuming that the corresponding turbulent normal stresses will be at least doubled, a magnitude of 1,500 dynes/em- can be estimated 6 mm downstream from the valve. Under pulsatile flow, the maximum turbulent normal stress measured 13 mm from the valve was 825 dynes/em', even though the magnitude may be correspondingly higher closer to the valve. Assuming a factor of 2, even at 13 mm from the valve, maximum turbulent shear stresses of about 400 dynes/em' can be expected downstream from the valve under pulsatile flow. Blackshear" has shown that red blood cells that adhere to foreign surfaces may be damaged by shear stresses of the order of 10 to 100 dynes/em', Sutera and Mehrjardi" have demonstrated that red blood cells lose their biconcavity when subjected to bulk turbulent shear stresses of about 500 dynes/em'. Even though the magnitude of turbulent stresses measured downstream from the bileaflet valves are smaller than those of other valve geometries, the magnitudes may still be enough to induce hemolysis as well as subhemolytic damage. The velocity measurements in the central flow orifice reported in this study suggested that flow reversal and separation can be expected in the region of the leaflet hinges for a substantial portion of the cardiac cycle. Previously reported studies on clinical experience with the bileaflet valves suggest that thrombus formation in the region of the hinges prevented the proper opening and closing of the leaflets." 9 Bowen, Tri, and Wortham" suggested that the thrombotic process was actually initiated earlier in the postoperative period at a time when anticoagulation was inadequate. Nunez, Iglesia, and Sotillo? have also reported thrombus formation at the hinge of one leaflet, affecting its motion. Our studies of flow reversal and separation in this region appear to correlate with the initiation of the thrombotic process. Despite the superior hemodynamic characteristics of the bileaflet prostheses, thrombus formation might prevent the proper movement of the leaflets without adequate anticoagulant therapy. Thanks are due to Dr. Bahram Khalighi and Mr. Reza Fatemi for help in obtaining data. REFERENCES Bruss KH, Reul H, Van Gilse J, Knott E: Pressure drop and velocity fields at four mechanical heart valve prosthe-
St. Jude Medical valve 749
ses. Bjork-Shiley standard, Bjork-Shiley concave-convex, Hall-Kaster and St. Jude Medical. Life Support Systems 1:3-22, 1983 2 Gombrich PP, Villafana MA, Palmquist WE: From concept to clinical. The St. Jude Medical bileaflet pyrolytic carbon cardiac valve, Prosthetic Heart Valves, AP Yoganathan, EC Harrison, WH Corcoran, eds., Proceedings of the Fourteenth Annual Meeting of the Association for the Advancement of Medical Instrumentation, 1979, pp 181-212 3 Emery RW, Nicoloff DM: St. Jude Medical cardiac valve prosthesis. J THORAC CARDIOVASC SURG 78:269-276 4 Gabby S, Yellin EL, Frishman WH, Frater R WM: In vitro hemodynamic comparison of St. Jude, Bjork-Shiley and Hall-Kaster Valves, Trans Am Soc Artif Intern Organs, 26:231-235,1980 5 Yoganathan AP, Chaux A, Gray RJ, DeRobertis M, Matloff JM: Flow characteristics of St. Jude prosthetic valve. An in vitro and in vivo study. Artif Organs 6:288-294, 1982 6 Yoganathan AP, Chaux A, Gray RJ, Woo Y-R, DeRobertis M, William FP, Matloff JM: Bileaflet, tilting disc and porcine aortic valve substitutes. In vitro hydrodynamic characteristics. J Am Coli Cardiol 3:313-320, 1984 7 Gray RJ, Chaux A, Matloff JM, DeRobertis M, Raymond M, Stewart M, Yoganathan AP: Bileaflet, tilting disc and porcine aortic valve substitutes. In vivo hydrodynamic characteristics. J 'Am Coli Cardiol 3:321-327, 1984 8 Bowen TE, Tri TB, Wortham DC: Thrombosis of a St. Jude Medical tricuspid prosthesis. A case report. J THORAC CARDIOVASC SURG 82:257-262, 1981 9 Nunez L, Iglesia S, Sotillo J: Entrapment of leaflet of St. Jude Medical cardiac valve prosthesis by miniscule thrombus. Report of two cases. Ann Thorac Surg 29:567-569, 1980 10 Chandran KB, Yearwood TL: Experimental study of physiological pulsatile flow in a curved tube. J Fluid Mech 111:59-85, 1981 11 Chandran KB, Khalighi B, Chen C-J, Falsetti HL, Yearwood TL, Hiratzka LF: Effect of valve orientation on flow development past aortic valve prostheses in a model human aorta. J THORAC CARDIOVASC SURG 85:893-901, 1983 12 Chandran KB, Cabell GN, Khalighi B, Chen C-J: Laser anemometry measurements of pulsatile flow past aortic valve prostheses. J Biomech 16:865-873, 1983 13 Chandran KB, Cabell GN, Khalighi B, Chen C-J: Pulsatile flow past aortic valve bioprostheses in a model human aorta. J Biomech 17:609-619, 1984 14 Blackshear P: Hemolysis and prosthetic surfaces, Chemistry of Biosurfaces, Vol 2, ML Hair, ed., New York, 1972, Marcel Dekker, Inc., pp 523-561 15 Sutera SP, Mehrjardi NH: Deformation and fragmentation of human RBC in turbulent shear flow. Biophys J 15: 1-10, 1975
CARDIOVASC SURG
89:743-749, 1985
Pulsatile flow past 81. Jude Medical bileaflet valve An in vitro study An in vitro hemodynamic study of the St. Jude Medical bileaflet aortic prosthesis was performed in a mock circulatory system simulating physiological pulsatile flow. The study included measurements of pressure drop across the valves, percent regurgitation, velocity, and turbulence in a model human aorta. The measurements indicated that pressure drop (mean systolic pressure drop of 6.2 mm Hg), percent regurgitation (10.15%), and turbulent normal stresses immediately downstream from the valve (825 dynes/cm~ were better than those with other prosthetic valvesand bioprostheses. The flow development in the aorta was not significantly affected by the orientation of the bileaflet valve in the root of the aorta. However, velocity measurements immediately downstream from the valves showed flow reversal and separation in the vicinity of the binge points of the leaflets where thrombus formation bas been previously reported.
Krishnan B. Chandran, D.Sc., Iowa City, Iowa
h e centrally occluding caged ball valve was introduced as a replacement for diseased aortic valves in 1960. Within a decade, with the introduction ofpyrolytic carbon discs, tilting disc valves gained popularity. More recently, a bileaflet valve geometry has been introduced into the market. The 81. Jude Medical bileaflet valve has two leaflets mounted in an orifice ring all coated with pyrolytic carbon. The leaflets open to about 85 degrees, and there are two major flow orifices and a minor central orifice. The valve has a low profile without any protecting guiding struts on the aortic side. In the closed position, the leaflets do not overlap and leave a ring slit between the leaflets and the orifice ring as well as a central slit where the two orifices meet.' In vitro hemodynamic evaluation of the relatively new geometry for valve prostheses has been previously reported.t' Gombrich, Villafana, and Palmquist/ compared the pressure gradients across the valve during steady and pulsatile flow, the flow dynamics in pulsatile From the Department of Biomedical Engineering, University of Iowa, Iowa City, Iowa Supported in part by Grant HL-26269 from the U.S. Public Health Service. Received for publication Jan. 30, 1984. Accepted for publication May 31, 1984. Address for reprints: K.B. Chandran, D.Sc., 1218 EB, Department of Biomedical Engineering, College of Engineering, The University of Iowa, Iowa City, Iowa 52242.
flow through qualitative flow visualization studies, and the calculated valve orifice areas between the bileaflet valve and other prosthetic and tissue valve. Their results suggested that the bileaflet valve had relatively lower pressure gradients than did other valve geometries, and computed valve orifice areas were larger for the bileaflet valves with comparable tissue anulus diameters. Bruss and associates! compared the pressure drop across the valves during steady flow for bileaflet and tilting disc valves and compared the velocity profiles downstream from the valves under pulsatile flow conditions with laser Doppler anemometry (LOA). However, they did not attempt to measure the fluctuating component of the velocity in their studies. Their results showed that the bileaflet valves had the lowest pressure gradient across the valve under steady flow conditions. These results have been further confirmed by Emery and Nicoloff" and Gabbay and associates.' The velocity profiles during pulsatile flow conditions by Bruss and co-workers! showed that the flow downstream to the bileaflet valve was relatively flat compared to the tilting disc valves and the wall shear stress could also be expected to be the least with the bileaflet valves. More recently, Yoganathan's group':" has performed in vitro pressure-drop studies and velocity measurements downstream from the St. Jude Medical valve under steady flow conditions. Gray and colleagues' have also discussed the in vivo hydrodynamic characteristics of bileaflet, tilting disc, and porcine valves. These results suggest that the 743
The Journal of
744
Chandran
Thoracic and Cardiovascular Surgery
Fig. 1. Orientation of the bileaflet valve with respect to the aortic root. A, Tilt axis in the plane of primary curvature. B, Tilt axis perpendicular to the plane of primary curvature. LCA, Left coronary artery. RCA, Right coronary artery.
bileaflet valve has better flow characteristics than the other valves. Despite the improved hemodynamic characteristics exhibited by the bileaflet valve in in vitro evaluation, thromboembolic complications have been reported in the clinical use of this prosthesis. 8,9 In this paper, a hemodynamic evaluation of the St. Jude Medical bileaflet prosthesis in vitro is reported. The measurements included ·pressure-drop studies, qualitive visualization of flow development past the prosthesis in a model human aorta, and quantitative measurement of velocity profiles and turbulent stresses performed with LDA. The results of the experimental study are compared with those obtained with other types of prosthetic valves as well as bioprostheses. Experimental procedure A St. Jude Medical bileaflet pyrolytic carbon valve with a tissue anulus diameter of 27 mm was used in this study. The valve was sutured to an acrylic valve mounting ring and incorporated into a mock circulatory system simulating physiological pulsatile flow. The mock circulatory system used in our laboratory has been described previously'? and was run at 72 beats/min with a systolic duration of 270 msec. The blood analog fluid used in these experiments was a glycerol solution (36% by volume in distilled water) whose viscosity coefficient and density at room temperature were representative of that of human blood at high rates of shear. Qualitative flow visualization studies were performed with illuminated, neutrally buoyant resin beads in a model human aorta as described by Chandran and associates. I I These studies were followed by quantitative measurements
Fig. 2. Flow visualization in the ascending aorta in early systole with the bileaflet valve. A, Valve oriented with the tilt axis in the plane of primary curvature (orientation A). B, Valve oriented with the tilt axis perpendicular to the plane of primary curvature (orientation B).
with LDA in a forward scattering mode at various cross sections in the ascending aorta and in the mid-arch region. The details of the measurements with LDA in our laboratory have been described elsewhere. 12 For these studies, the bileaflet valve was mounted in two different orientations with respect to the root of the aorta, so that the effect of orientation on the flow development could be analyzed. In orientation A, the tilt axis was in the plane of the primary curvature of the aorta and in orientation B, the tilt axis was perpendicular to the plane of primary curvature, as shown in Fig. I. The pressure-drop measurements across the bileaflet
Volume 89
St. Jude Medical valve 745
Number 5 May. 1985
valve were obtained under physiological pulsatile flow conditions with a Validyne P35d differential pressure transducer. For the pressure-drop studies, an axisymmetric valvular chamber" was incorporated into the mock circulatory system replacing the model human aorta. The axisymmetric flow chamber had pressure taps both upstream and downstream from the valve for the pressure-drop measurements. This chamber was also used to obtain velocity measurements 21 mm downstream from the seat of the valve, whereas in the aorta, data could be obtained no closer to the valve than 28 mm in the experimental setup used in these studies. The measurement 21 mm downstream from the seat of the valve is approximately 13 mm downstream from the valve, whereas Yoganathan and associates'" made the measurements 6 mm downstream from the valve under steady flow conditions. Results Flow visualization study. Flow visualization study was performed by recording the motion of the illuminated resin beads in the aortic flow chamber. Highspeed movies at 100 frames/sec were obtained and the motion of the particles was analyzed with a motion analyzer projector. Also, 35 mm photographs were obtained with a 35 mm SLR camera driven by a motor driveat 3 frames/sec. The motor drive was triggered by the electronic pulser so that the picture could be timed in reference to the cardiac cycle. Fig. 2, A and B shows typical flow visualization pictures in early systole for orientations A and B, respectively. In Fig. 2, A, the valve is oriented such that the minor central flow orifice is in the plane of the picture. In early systole, a uniform forward motion is observed in the ascending aorta. In late systole, this uniform flow is strongly affected by the secondary motion because of a large amount of flow through the major flow orifices near the lateral walls. In early diastole, a flow reversal is initiated along the inner wall of curvature of the aorta (posterior wall), which is soon extended across the cross section in the mid-arch region. In orientation B, with the central flow orifice perpendicular to the plane of the picture, relatively fast fluid motion is observed in the region of major flow orifices, with very little motion in the core region downstream from the central minor orifice. However, as the flow develops, a more uniform forward motion is observed in the ascending aorta. In early diastole, a flow reversal is observed along the inner wall of curvature. A slight forward motion is observed along the outer wall of curvature through a large part of diastole, with a resulting clockwise vortex motion in the plane of the picture in the aortic chamber. In contrast to the study on
20.0 r - - - - - - - - - - - - - - - - ,
[
"'"<
~
3 LL
U 0::
I-
w
15.0
10.0
5.0
:2
::J -.J
'~
0.0 1-------t---f"r----::::;iII"'c---.,
- 5.0
L..-_ _---l.
0.00
0.21
---lIoL-_ _.....L.
0.42 TIME (sec)
0.63
....l
0.84
Fig. 3. A typical instantaneous flow rate signal in a cardiac cycle.
the effect of valve orientation of tilting disc valves,II in neither orientation did the flow development past bileaflet valves appear to produce regions of relative stasis in the ascending aorta or in -the mid-arch region. Pressure-drop and velocity profiles in the axisymmetric chamber. A typical instantaneous flow rate profile in a cardiac cycle as recorded from a Carolina electromagnetic flowmeter is shown in Fig. 3. The cardiac output and percent regurgitation (ratio of retrograde flow over the net positive flow) can be determined by computing the area under the curve for positive and negative flow. In these experiments, the cardiac output was maintained to be about 6 L/m and the percent regurgitation was computed to be 10.15%. This value compared favorably with those for other valve prostheses evaluated in our laboratory. The percent regurgitation was in the range of 10% to 13% for tilting disc valves (Bjork-Shiley and Hall-Kaster), 3% to 10% for caged ball valves (Starr-Edwards), and 3% to 7% for bioprostheses (porcine and pericardial). The peak and the mean systolic pressure drop across the valve were determined with the differential pressure transducer. A mean systolic pressure drop of 6.2 mm Hg was measured for the S1. Jude Medical valve, and the value ranged between 6 and 12 mm Hg for the prosthetic valves. Our results showing one of the lowest pressure drops with the bileaflet valves agreed with those studied reported earlier. I. 2 In the axisymmetric aortic chamber, the velocity profiles 21 mm downstream from the seat of the bileaflet valve were obtained for two different orienta-
The Journalof Thoracic and Cardiovascular Surgery
7 4 6 Chandran
125.0 , - - - - - - - - - - - - - - - - - - - - ,
125.0 , - - - - - - - - - - - - - - - - - - ,
100.0
100.0
E 75.0 VI
~ 75.0 E (J
(J
~ 50.0
50.0
~
U
u
o
o
....J
....J
w 25.0
~ 25.0
>
....J
....J
X
X 0.0
0.0
-25.0 .... --.rL~ -50.0 '--
-25.0 J-
-'--
---'-
-1.0 -0.5 0.0 0.5 NEAR WALL NON-DIMENSIONAL RADIUS
---'
1.0 FAR WALL
Fig. 4. Velocity profiles 21 mm downstream from the St. Jude Medical valve perpendicular to the tilt axis. D = 40 msec after valve opening; V = 160 msec after valve opening; o = 210 rnsec after valve opening; • = 400 rnsec after valve opening.
tions. In the first orientation, the velocity measurements were made by transversing the measuring volume of the LDA perpendicular to the tilt axis of the leaflets. Hence, the velocity measurements were obtained in the two major flow orifices as well as in the central minor flow orifice, In the second orientation, the LDA traverse was parallel to the tilt axis such that the velocity profile in the central flow orifice was obtained. Fig. 4 shows the velocity profiles at different times in the pulsatile flow cycle for the first orientation. In early systole (40 msec after valve opening), a forward flow throughout the cross section is observed, and by peak systole (160 msec), a three-peaked velocity profile in the core region is observed with region of flow reversal and vortex formation in the valvular sinus region. The three-peaked velocity profile corresponding to the three flow regions of the bileaflet valve is very similar to that reported by Bruss et al.' In early diastole, the vortices in the sinus region are still observed while the flow in the core region is reduced and the fluid motion dies out toward the rest of the cardiac cycle in the pulse duplicator. Fig. 5 compares the velocity profile for the two orientations of the valve in peak systole (160 msec). In the orientation parallel to the tilt axis, when the velocity measurement is in the central flow orifice, a single-
-50.0 '--
...l..-
-L
----'-
---'
-1.0 - 0.5 0.0 0.5 1.0 NEAR WALL FAR WALL NON-DIMENSIONAL RADIUS
Fig. 5. Velocity profiles 21 mm downstream from the St. Jude Medical valve in peak systole (160 msec). 0 = Perpendicular to the tilt axis; D = parallel to the tilt axis. peaked velocity profile is observed in the core region with relative stasis in the periphery near the leaflet housing. This region of very little fluid motion is observed for a large portion of the cardiac cycle. Should depositions accumulate in this region, the leaflet motion may be obstructed. Fig. 6 compares the velocity profiles in peak systolic in the axisymmetric chamber for the St. Jude Medical valve with those of Bjork-Shiley convexo-concave tilting disc valve (Model 27 MBRM), a Carpentier-Edwards porcine valve (Model 2625), and an Ionescu-Shiley low-profile pericardial valve. The tilting disc valve with the minor and major flow regions yields a jet-like flow in the major flow orifice with relatively large wall shear stresses and a vortex formation in the minor flow region." With the porcine valve, which has a relatively smaller valve orifice area than the pericardial valve, a strong central core jet is observed with a large vortex formation between the commissure and the valve support, which protrudes into the lumen. The velocity profile for the pericardial valve shows a uniform core flow and a vortex formation between the leaflets and the sinus. The tilting disc valve, along with the caged ball valve, yields flow with relatively large wall shear stresses. The flow past the porcine valve shows a strong central jet with high turbulent stresses." Our measurements showed that the turbulent normal stresses downstream to the valves are about 1,800 to 2,600 dynes/em' for the
Volume 89 Number 5
St. Jude Medical valve
747
May, 1985
porcine valve, about 1,200 to 1,600 dynes/em' for the caged ball, tilting disc, and the pericardial valves, and about 825 dynes/em' for the bileaflet valve. Measurement of turbulent shear stress, which has been directly linked to the damage to formed elements in blood,14,15 requires simultaneous measurement of the axial and radial components of velocity. With a single-channel laser system in our laboratory, it was not possible to measure the turbulent shear stress. Even though the St. Jude Medical valve has one of the lowest pressure gradients of all valve prostheses, along with the relatively smaller turbulent stresses, thromboembolic complications are still reported with this valve.8,9 Velocity profiles in the aorta. Velocity profiles were measured in several cross sections in the ascending aorta, mid-arch region, and in the brachiocephalic artery as shown schematically in Fig. 7. The LDA traverses were made from the right lateral to the left lateral wall in plane AA (28 mm from the seat of the valve), plane BB (48 mm from the seat of the valve), plane CC, and plane DD. Measurements were obtained with the valves mounted in both the orientations described in the flow visualization studies (Fig. 1). Figs. 8 and 9 show the velocity profiles in peak systole (160 msec after valve opening) at cross sections AA and BB, respectively, for both orientations A and B. Even in section AA, 28 mm downstream from the seat of the valve, the velocity profiles are relatively flat. The flow reversal observed near the sinus region or near the leaflet hinges described earlier (Fig. 5) are not apparent in this cross section. In section BB, both the velocity profiles show a tendency to skew toward the left lateral wall, which is the inner wall of the tertiary curvature shown in Fig. 7. However, the flow development in the ascending aorta or in the mid-arch region was not observed to change dramatically with the change in orientation of the bileaflet valve. In the mid-arch region also, the velocity profiles were observed to be relatively flat for both the orientations; in diastole, a flow reversal was observed throughout the mid-arch region with the bileaflet valves. Relatively flat velocity profiles were observed in the brachiocephalic arterial branch (cross section DD) for both orientations of the valve. Discussion The in vitro measurements of pulsatile flow past the St. Jude Medical bileaflet valve suggests that the flow characteristics are comparable, if not superior, to those of prosthetic valves of caged ball and tilting disc geometry. The mean systolic pressure gradient and the percent regurgitation with the bileaflet valves compared favorably with other prosthetic valves as well as biopros-
150,0
120,0
'"
<,
E
90,0
u
>-
t<.)
0
60,0
--.J W
>
--.J
« x «
30.0
0.0
-30,0 ' - - - - _ - - ' - -1.0 - 0, 5 NEAR WALL
---'-0.0
-----'-0.5
-.1
1.0 FAR WALL
NON-DIMENSIONAL RADIUS
Fig. 6. Velocity profiles 21 mm downstream to the valve seat in peak systole. 0 = St. Jude Medical valve; • = Bjork-Shiley convexo-concave valve; 'V = Carpentier-Edwards procine valve; 0 = Ionescu-Shiley low-profile valve,
theses. The measured velocity profiles with the bileaflet valves showed a centralized flow with relatively lower wall shear stresses than with the caged ball and tilting disc valves. The measured turbulent axial stresses were also relatively lower with the bileaflet values than with the other valve prostheses. Flow visualization studies complemented by quantitative velocity measurements in the model human aorta suggested that the orientation of the bileaflet valve at the root of the aorta did not affect the flow development in the ascending aorta. In contrast, with the tilting disc valves, the flow development strongly depended on the orientation of the major flow orifice. I I A comparison of the velocity profiles under pulsatile flow reported in this work with those reported under steady flow" show several interesting features, As pointed out earlier, pulsatile flow measurements were obtained in this study 13 mm downstream from the valve at a peak instantaneous flow rate of 18 Lzrn, whereas the steady flow measurements" were obtained 6 mm downstream from the valve at a flow rate of 25 L/min. The velocity profiles are similar in shape in both orientations for steady and pulsatile flow. However, under pulsatile flow the flow reversal is stronger in the valvular sinus area, with a negative velocity of about 25 em/sec compared to 8 em/sec under steady flow, even though the pulatile flow measurements were made
The Journal of
7 4 8 Chandran
Thoracic and Cardiovascular Surgery
OUTER WALL OF PRIMARY CURVATURE (GREATER CURVE) OUTER WALL OF SECONDARY CURVATURE
+---- DIRECTION OF THE
VELOCITY MEASUREMENT
DIRECTION OF FLOW
Fig. 7. Schema of the model human aorta and the cross sections in which velocity measurements were obtained. 125.0
r------------------....,
100.0 , - - - - - - - - - - - - - - - - - - -
100.0
e'" o
~ 75.0 E
7 5 .O
u
>-
~ 50.0
fooU
o
:l 50.0
~ 25.0
>
-l
-l
u
-l
sx
UJ
0.0
X
-25.0
-50.0 ' - - - - - - - ' - - - - - - - ' ' - - - - - - ' - - - - - - - ' -1.0 -0.5 0.0 0.5 1.0 RIGHT LEFT LATERAL WALL LATERAL WALL
NON-DIMENSIONAL RADIUS
Fig. 8. Velocity profiles in peak systole (160 msec after valve opening) in cross section AA in the ascending aorta. 0 = Orientation A; 0 = orientation B.
0.0 O - - - - - - ' - - - - - - - - ' - - - - - ' - - - - - - a -1.0 - 0.5 0.0 0.5 1.0 RIGHT LEFT LATERAL WALL LATERAL WALL
NON-DIMENSIONAL RADIUS
Fig. 9. Velocity profiles in peak systole (120 msec after valve opening) in cross section BB in the ascending aorta. 0 = Orientation A; 0 = orientation B.
Volume 89 Number 5 May, 1985
further downstream from the valve. The peaks and valleys in the velocity profile in the core region corresponding to the three flow orifices are more prominent in steady flow closer to the valve. Yoganathan and associates" reported maximum turbulent shear stresses of 750 dynes/em- under steady flow. Assuming that the corresponding turbulent normal stresses will be at least doubled, a magnitude of 1,500 dynes/em- can be estimated 6 mm downstream from the valve. Under pulsatile flow, the maximum turbulent normal stress measured 13 mm from the valve was 825 dynes/em', even though the magnitude may be correspondingly higher closer to the valve. Assuming a factor of 2, even at 13 mm from the valve, maximum turbulent shear stresses of about 400 dynes/em' can be expected downstream from the valve under pulsatile flow. Blackshear" has shown that red blood cells that adhere to foreign surfaces may be damaged by shear stresses of the order of 10 to 100 dynes/em', Sutera and Mehrjardi" have demonstrated that red blood cells lose their biconcavity when subjected to bulk turbulent shear stresses of about 500 dynes/em'. Even though the magnitude of turbulent stresses measured downstream from the bileaflet valves are smaller than those of other valve geometries, the magnitudes may still be enough to induce hemolysis as well as subhemolytic damage. The velocity measurements in the central flow orifice reported in this study suggested that flow reversal and separation can be expected in the region of the leaflet hinges for a substantial portion of the cardiac cycle. Previously reported studies on clinical experience with the bileaflet valves suggest that thrombus formation in the region of the hinges prevented the proper opening and closing of the leaflets." 9 Bowen, Tri, and Wortham" suggested that the thrombotic process was actually initiated earlier in the postoperative period at a time when anticoagulation was inadequate. Nunez, Iglesia, and Sotillo? have also reported thrombus formation at the hinge of one leaflet, affecting its motion. Our studies of flow reversal and separation in this region appear to correlate with the initiation of the thrombotic process. Despite the superior hemodynamic characteristics of the bileaflet prostheses, thrombus formation might prevent the proper movement of the leaflets without adequate anticoagulant therapy. Thanks are due to Dr. Bahram Khalighi and Mr. Reza Fatemi for help in obtaining data. REFERENCES Bruss KH, Reul H, Van Gilse J, Knott E: Pressure drop and velocity fields at four mechanical heart valve prosthe-
St. Jude Medical valve 749
ses. Bjork-Shiley standard, Bjork-Shiley concave-convex, Hall-Kaster and St. Jude Medical. Life Support Systems 1:3-22, 1983 2 Gombrich PP, Villafana MA, Palmquist WE: From concept to clinical. The St. Jude Medical bileaflet pyrolytic carbon cardiac valve, Prosthetic Heart Valves, AP Yoganathan, EC Harrison, WH Corcoran, eds., Proceedings of the Fourteenth Annual Meeting of the Association for the Advancement of Medical Instrumentation, 1979, pp 181-212 3 Emery RW, Nicoloff DM: St. Jude Medical cardiac valve prosthesis. J THORAC CARDIOVASC SURG 78:269-276 4 Gabby S, Yellin EL, Frishman WH, Frater R WM: In vitro hemodynamic comparison of St. Jude, Bjork-Shiley and Hall-Kaster Valves, Trans Am Soc Artif Intern Organs, 26:231-235,1980 5 Yoganathan AP, Chaux A, Gray RJ, DeRobertis M, Matloff JM: Flow characteristics of St. Jude prosthetic valve. An in vitro and in vivo study. Artif Organs 6:288-294, 1982 6 Yoganathan AP, Chaux A, Gray RJ, Woo Y-R, DeRobertis M, William FP, Matloff JM: Bileaflet, tilting disc and porcine aortic valve substitutes. In vitro hydrodynamic characteristics. J Am Coli Cardiol 3:313-320, 1984 7 Gray RJ, Chaux A, Matloff JM, DeRobertis M, Raymond M, Stewart M, Yoganathan AP: Bileaflet, tilting disc and porcine aortic valve substitutes. In vivo hydrodynamic characteristics. J 'Am Coli Cardiol 3:321-327, 1984 8 Bowen TE, Tri TB, Wortham DC: Thrombosis of a St. Jude Medical tricuspid prosthesis. A case report. J THORAC CARDIOVASC SURG 82:257-262, 1981 9 Nunez L, Iglesia S, Sotillo J: Entrapment of leaflet of St. Jude Medical cardiac valve prosthesis by miniscule thrombus. Report of two cases. Ann Thorac Surg 29:567-569, 1980 10 Chandran KB, Yearwood TL: Experimental study of physiological pulsatile flow in a curved tube. J Fluid Mech 111:59-85, 1981 11 Chandran KB, Khalighi B, Chen C-J, Falsetti HL, Yearwood TL, Hiratzka LF: Effect of valve orientation on flow development past aortic valve prostheses in a model human aorta. J THORAC CARDIOVASC SURG 85:893-901, 1983 12 Chandran KB, Cabell GN, Khalighi B, Chen C-J: Laser anemometry measurements of pulsatile flow past aortic valve prostheses. J Biomech 16:865-873, 1983 13 Chandran KB, Cabell GN, Khalighi B, Chen C-J: Pulsatile flow past aortic valve bioprostheses in a model human aorta. J Biomech 17:609-619, 1984 14 Blackshear P: Hemolysis and prosthetic surfaces, Chemistry of Biosurfaces, Vol 2, ML Hair, ed., New York, 1972, Marcel Dekker, Inc., pp 523-561 15 Sutera SP, Mehrjardi NH: Deformation and fragmentation of human RBC in turbulent shear flow. Biophys J 15: 1-10, 1975