- Email: [email protected]
Leaflet fracture in Edwards-Duromedics bileaflet valves
J
THoRAc CARDIOVASC SURG
1989;97:90-4
Leaflet fracture in Edwards-Duromedics bileaflet valves Two cases of leaflet fracture in the Edwards-Duromedics valve at 36 and 38 months after implantation are reported. Both patients were immediately reoperated on and recovered weD. In one valve an older housing fracture with partial tissue ingrowth was noted beside a recent transverse leaflet fracture. In the othervalve theleafletwas fractured nearthe pivot mechanism. AU larger embolized parts were detected in the iliac artery region by computed tomographic scan and were subsequently removed. Problems in diagnosis and the importance of immediate reoperation, even without exact diagnosis, are discussed. Technical evaluation of the valve revealed crack growth and arrest, giving evidence of fatigue fracture. Scanning electron microscopic examination revealed several areas of pitting and erosion. Although the exact cause of mechanical disruption remains speculative, pyrolytic carbon seems to have the characteristic of fatigue fracture as weD as erosion damage. A connection between the two might exist
W. Klepetko, MD,a A. Moritz, MD,a J. Mlczoch, MD,b H. Schurawitzki," E. Domanig, MD,a and E. Wolner, MD,a Vienna. Austria
Mechanical disruption has been described in almost every type of artificial valve.!" The Edwards-Duromedies bileaflet valve (Baxter Edwards Divisions, Baxter Healthcare Corporation, Irvine, Calif.) has now been used for 4 years at our institution, and we5• 6 have reported satisfying clinical results and no instances of mechanical failure or disruption. Recently, two of our patients have experienced leaflet fracture. Complete clinical workup of these two events has been done at our institution, and both valves were investigated by the manufacturer for underlying mechanical reasons for failure. We believe these events to be the first reported cases of leaflet fracture in Edwards-Duromedies valves in the open literature.
Case reports CASE 1. A 41-year-old man underwent mitral valve replacement in October 1983 because of mixed rheumatic mitral valve failure. An Edwards-Duromedics 31 mm mitral valve
From the Second Surgical Department,' University of Vienna (Head: Prof. E. Wolner), the Cardiological Department," University of Vienna (Head: Prof. F. Kaindl), and the Department of Radiology,' University of Vienna (Head: Prof. H. Pokiesser), Vienna, Austria. Received for publication Feb. 7, 1988. Accepted for publication June 13, 1988. Address for reprints: Walter Klepetko, MD, Second Surgical Department, University of Vienna, Spitalgasse 23, A-I090 Vienna, Austria.
90
prosthesis was implanted and the postoperative course was uneventful. At follow-up of 3, 6, 12, and 24 months the patient was in New York Heart Association class I. He had no complaints and his anticoagulation program was well controlled. The last echocardiographic examination was performed in May 1986. It revealed slightly reduced left ventricular function and normal function of the prosthesis. In December 1986 the patient had sudden severe dyspnea while running up several stairs. He was then admitted to the department of cardiology because of pulmonary edema. In the acute phase the echocardiogram revealed good left ventricular function, but the function of the mitral valve prosthesis could not be determined because of tachycardia and tachypnea. After stabilization of the patient, the opening of only one leaflet could be detected; opening dysfunction of the other leaflet was, therefore, suspected. Doppler investigation revealed no significant mitral regurgitation, but the mitral flow velocity was increased to 3 msec, and the diagnosis of acute valve dysfunction was made. The patient was referred for operation immediately thereafter. At operation, the mitral valve was found to be well healed in, but one leaflet was missing; the remaining leaflet was moving freely. There were no obvious signs of tissue overgrowth or thrombus formation. The missing leaflet could not be detected within the cardiac cavity. The Edwards-Duromedies valve was excised and a Bjork-Shiley Monostrut 31 nun mitral valve (Shiley Incorporated, Irvine, Calif.) was implanted. The patient recovered quickly. On the fourth postoperative day ultrasound investigation of the great vessels revealed two foreign bodies in the iliac arteries, one on each side. This finding was confirmed by computed tomography (Fig. I). Both iliac arteries were explored by a retroperitoneal approach and two leaflet pieces were extracted. Further recovery of the patient was uneventful. Investigation of the leaflet showed a transverse fracture
Volume 97 Number 1
Edwards-Duromedics bileaflet valves
January 1989
9I
Fig. 1. Computed tomographic scan of the iliac arteries showing an embolized leaflet.
dividing it into two pieces of approximately one third and two thirds. On the excised valve, the highest part of the pyrolytic carbon housing was missing around about half the circumference on the side of the escaped leaflet. The outer metal ring, which was overgrown with a thin layer of tissue, was visible (Fig. 2). CASE 2. A 15-year-old boy was operated on because of rheumatic double valve disease in January 1984. He underwent double valve replacement with Edwards-Duromedics valves: 23 mm aortic and 31 mm mitral. Recovery was uneventful and he was discharged, but he refused any followup control studies because of lack of complaints. Thirty-six months after the operation, he suddenly had severe dyspnea during a walking tour and his condition rapidly deteriorated. On admission to the cardiological department, he was in acute pulmonary edema with severe tachypnea and tachycardia. After his condition was stabilized with medical treatment, echocardiograms were obtained. Left ventricular function was found to be good. The function of the aortic and mitral prostheses, however, could not be determined. Doppler investigation suggested light aortic regurgitation . but no mitral regurgitation could be shown. However, aortography gave no evidence of aortic insufficiency. Right heart catheterization revealed a v-wave of 80 mm Hg of pulmonary capillary wedge pressure. On fluoroscopic examination, abnormal movement of the mitral valve housing within the anulus was seen. which indicated unusual forces acting on the housing. Therefore, the patient was referred for exploration of the mitral valve. At reoperation, one leaflet of the mitral valve was missing. There was neither tissue overgrowth nor signs of thrombosis, but two small perivalvular leaks of 3 and 5 mm diameter were detected. The mitral valve was excised and a 29 mm BjorkShiley Monostrut valve was implanted . On the fifth postoperative day, ultrasound investigation of the iliac arteries was performed, but no foreign body could bedetected. By means of computed tomography, the leaflet was imaged in the left common iliac artery. It was then removed by a retroperitoneal approach. Further recovery was uneventful. Investigation of the excised valve revealed a normal valve
Fig. 2. Explanted valve from the first 'patient together with the two pieces of the escaped leaflet. Tissue ingrowth can be noted on the housing fracture surface.
housing and normal motion of the remaining leaflet. The escaped leaflet was fractured through one of the pivots (Fig. 3).
In vitro examination Both valves were returned to the manufacturer and subjected to a series of examinations in accordance with a standard protocol to determine facts about the events. The valves were photographed by macrophotographic and microphotographic techniques. X-ray films were made to examine subsurface structures, and a pathologic examination of the tissue was done. After these examinations, microscopic visual exams at 40X magnification were done to look for any unusual features on the valve. The suture rings were then removed and the valves were cleaned of all biologic debris. Scanning
92
The Journal of Thoracic and Cardiovascular Surgery
Klepetko et al.
• Fig. 3. Explanted valve from the second patient together with the major portion of the escaped leaflet. The fracture line can be seen at the upper pivot region.
Fig. 4. SEM of housing fracture surface in valve I, illustrating secondary cracking into the pivot region and the crack growth and arrest phenomenon.
electron microscopy (SEM) was performed and appropriate photographs at set magnifications were taken. Wherever feasible the valves were remeasured for compliance with original specifications and then disassembled. SEM examination was again performed after disassembly. Other nondestructive tests to determine local hardness, silicone content, or other properties were conducted. Valve 1. The pathologic examination detected tissue ingrowth over portions of the housing fracture surface. This observation supported the conclusion that this fracture had occurred long before the leaflet escaped. From light microscopic examination, there was evidence of pitting in the area of the seating surface of the housing, as well as in the pivot area of the remaining
Fig. 5. SEM of leaflet fracture surfacein valve I. The pitting can be seen on the fracture line near the edge of the leaflet.
leaflet. The escaped leaflet was in two pieces; the fracture had occurred from the seating edge at about 60 degrees from one pivot through the center of the rear of the leaflet. A pit approx imately 0.4 mm in diameter was noticeable along the fracture line. SEM examination demonstrated that the housing fracture had originated centrally and had grown circumferentially toward each pivot (Fig.A). The section fractured out at the transition from the curved seating area to the pivot flats. The fracture line was above the leaflet-housing contact line toward the outflow side of the valve. Crack growth and arrest marks in both fracture surfaces of the leaflet were demonstrated as well. The aforementioned pit at the fracture line was confirmed to blend gradually into the normal surface and abruptly into the fracture surface (Fig. 5), which suggested that the fracture originated at the pitting area. Additional smaller pits occurred on the surface of both leaflets and in the pivot ball area . Valve 2. Light microscopic studies revealed heavy chipping and pitting in the pivot slot area of the housing, both in the side of the escaped leaflet and in the opposing side. A large porous area was noted in the base of the pivot slot of the escaped leaflet opposite the side where pivot ball fracture occurred. Many porous areas around the circumference of the seating lip were also seen. Pitting was noted around the seating rim and in the pivot ball area opposite the fractured portion of the escaped leaflet (Fig. 6). The leaflet that remained in the valve had corresponding damage to its rear surface and some pitting around the seating rim, which gave evidence of interference damage. SEM examination showed a complex fracture surface through the pivot ball area (Fig. 7), the possible origin being in a slightly pitted area. The fracture surface again had evidence of
Volume 97 Number 1 January 1989
Edwards-Duromedics bileaftet valves
93
Fig. 6. SEM of housing damage at the pivot slots on the side of the pivot ball fracture in valve 2, demonstrating heavy pitting and chipping.
Fig. 7. SEM of pivot fracture surface in valve 2, showing the crack growth and arrest phenomenon.
crack growth and arrest. There was evidence of interference on the rear surfaces of the leaflets. Both pitting and mechanical damage existed on this valve, and it appeared that a combination of flow and mechanical factors were at work in this fracture, Examination of all manufacturing records revealed no abnormalities and established no trends in these failures. On remeasurement of the valves, all dimensions were found to fulfill specifications . Measurements of local hardness, silicone content, and other mechanical properties were all found to be within the normal range.
artificial valve is to establish the diagnosis and initiate appropriate treatment. The acute onset of severe symptoms in these patients is highly suggestive of mechanical dysfunction, as opposed to a more gradual onset, which might result from valve thrombosis. However, confirming this suspicion is difficult, because the patients are in unstable distress, making it technically difficult to do clinical investigations. In both of our patients, neither M-mode nor Doppler echocardiography led to conclusive results. The physical examinations, electrocardiograms, and plane thoracic x-ray films reflect only the patient's condition and not the condition of the valve, but they do have the value of excluding myocardial infarction. Only the fluoroscopic examination in one patient gave evidence of disturbance of valve function. However, it still was not possible to establish a conclusive diagnosis in either case. Therefore, the key to the survival of these two patients was immediate reoperation without an exact diagnosis. On reoperation, the valve obviously has to be replaced and the cardiac cavity explored for the missing pieces; however, it is unlikely that they will be found in the heart, because they usually embolize . This is true even in double valve replacement, as the experience with the second patient shows, in whom a 31 mm leaflet passed through a 23 mm aortic valve, without disrupting its function. Embolized leaflets have not been reported to create urgent problems; therefore, the patient's condition can be allowed to stabilize before attempts to locate and retrieve the parts are made. In our patients, we tried to find the leaflets on the fourth and fifth postoperative days. This was done by computed tomographic scan and ultrasonic imaging, which proved to be complementary
Discussion Several clinical and technological aspects of these two events are remarkable. To our knowledge this is the first report of a leaflet fracture in the Edwards-Duromedics bileaflet valve, which has been in clinical use since 1982. Mechanical disruption in other bileaflet valves has been described by others . I. 7. 8 As compared to occluder escape in single disc valves? or ball valves,'? this event in bileaflet valves normally offers a better chance for adequate treatment and patient survival, since the escape of only one leaflet still leaves some residual function of the valve. Disc escape in single disc valves has been reported to result in sudden death," but single leaflet escape in a bileaflet valve has not caused sudden death in our experience. This offers valuable time , especially in patients who are in good condition before the event. Whether this situation also pertains to patients in higher New York Heart Association classes cannot be answered by our experience. The essential factor in acute dysfunction of an
94
The Journal of Thoracic and Cardiovascular Surgery
Klepetko et al.
techniques. However, one must be aware that more than one foreign body can exist, and the search for multiple pieces should be conducted as accurately as possible. The usual location for the embolized parts is the iliac artery region, where we found the leaflet components. The small fractured pivot ball and the missing housing part were not detected and probably have embolized into smaller peripheral arteries without creating symptoms. Embolized discs of the Bjork-Shiley valve, which are larger than a leaflet of a bileaflet valve, have generally been found in the abdominal aorta.': II It is remarkable that the greatest number of reported cases of mechanical disruptions have happened in mitral valves.lv" This seems to be caused by particularly strong mechanical forces acting on these valves. From the technical examinations, the nature of the fracture in both valves seems to be a fatigue fracture, as evidenced by crack growth and arrest. In the first patient, the earlier housing fracture led to tissue ingrowth, which in our opinion may have resulted in uneven closure of the leaflet. The cause of the housing fracture remains unclear. In the operating protocol, the surgeon mentioned the removal of a calcified particle through the valve orifice after the implantation, which speculatively might have had a causal relation to the housing fracture. Another possibility for microcrack initiation could arise from the manufacturing process when the leaflets are installed or the stiffening ring is placed onto the housing. The eventual fracture of the leaflet, probably starting from one of the pitted spots, may have resulted from uneven and unusual closing forces over a long period of time. In the second valve, evidence for interference between the two leaflets was found. This was seen by unusual damage in corresponding areas of the hinge mechanism and the leaflets. Again, crack growth and arrest support fatigue as the cause of the fracture. The cause of the pitting at various locations on the carbon surface of the valve remains unclear. It could result from local mechanical forces arising from instruments or interference, from local flow conditions that could create a jetting mechanism, from local variations in material properties, and finally from local arching created by defibrillation shocks. All these theories result from in vitro tests and observations that are currently in progress at the manufacturer, and evidence suggests that each of these theories has possible importance under certain circumstances. These findings underline the importance of careful manufacturing and proved quality control for all artificial valves to avoid potential release of inadequately qualified products. It also points out the necessity of careful handling of pyrolytic carbon valves and the need for awareness on the part of the surgeon that even
unnoticed damage can result in late mechanical disruption when combined with unusually high stresses in individual patients. In addition, pyrolytic carbon seems to have qualities that are not yet fully described. For instance, fatigue is an unknown phenomenon in this material, and erosion, to our knowledge, has been mentioned in only one other published report. 13 We wish to thank the Research and Development staff of Baxter Edwards CVS Division, Irvine, California, for their assistance in conducting the examinations of the explanted prostheses. REFERENCES I. Odell JA, Durandt J, Shama DM, Vythilingum S, Spontaneous embolization of a St. Jude prosthetic mitral valve leaflet. Ann Thorac Surg 1985;39:569-72. 2. Berry BE, Sheedy PF, McGoon DC. Diagnosis and management of aortic poppet embolism. Ann Thorac Surg 1974;17:504-9. 3. Davis PK, Myers JL, Pennock JL, Thiele BL. Strut fracture and disc embolization in Bjork-Shiley mitral valve prostheses: diagnosis and management. Ann Thorac Surg 1985;40:65-8. 4. Roberts AK, Lambert CJ, Mitchel BF. Embolization of disc occluder of a Wada-Cutter mitral prosthesis with survival. Ann Thorac Surg 1976;21:361-4. 5. Moritz A, Klepetko W, Grabenwoeger F, et al. Two-years experience with the Duromedics bileaflet valve prosthesis, Texas Heart Inst J 1985;12:315-22. 6. Klepetko W, Moritz A, Khunl-Brady G, et al. Implantation of the Duromedics valve prosthesis in 400 patients. Ann Thorac Surg 1987;44:303-9. 7. Hasse J. Escaped leaflet in a St. Jude Medical mitral prosthesis. In: DeBakey ME, ed. Advances in cardiac valves: clinical perspectives. New York: Yorke Medical Books, 1983:115-23. 8. Hjelms E. Escape of a leaflet from a St. Jude Medical prosthesis in the mitral position. Thorac Cardiovasc Surg 1983;31:310-2. 9. Bonnabeau RC Jr, Lillehei CWo Mechanical "ball fracture" in Starr-Edwards prosthetic valves. J THORAC CARDlOVASC SURG 1968;56:258-64. 10. Lindblom D, Bjork YO, Semb BK. Mechanical failure of the Bjork-Shiley valve: incidence, clinical presentation, and management. J THORAC CARDIOVASC SURG 1986;92:894-907. 11. Larrieu AJ, Puglia E, Allen P. Strut fracture and disc embolization of a Bjork-Shiley mitral valve prosthesis: localization of embolized disc by computerized axial tomography. Ann Thorac Surg 1982;34:192-5. 12. Marbarger JP, Clark RE. The clinical life history of explanted prosthetic heart valves. Ann Thorac Surg 1982;34:22-33. 13. Dobrova NB, lofis NA, Kozyrkin HI, Agafonov AV, Zaretskii YV, Kevorkova RA. Analysis of causes for wear and breakdown of elements in artificial cardiac valves of the disk type. Med Tekh 1987;2:8-12.
THoRAc CARDIOVASC SURG
1989;97:90-4
Leaflet fracture in Edwards-Duromedics bileaflet valves Two cases of leaflet fracture in the Edwards-Duromedics valve at 36 and 38 months after implantation are reported. Both patients were immediately reoperated on and recovered weD. In one valve an older housing fracture with partial tissue ingrowth was noted beside a recent transverse leaflet fracture. In the othervalve theleafletwas fractured nearthe pivot mechanism. AU larger embolized parts were detected in the iliac artery region by computed tomographic scan and were subsequently removed. Problems in diagnosis and the importance of immediate reoperation, even without exact diagnosis, are discussed. Technical evaluation of the valve revealed crack growth and arrest, giving evidence of fatigue fracture. Scanning electron microscopic examination revealed several areas of pitting and erosion. Although the exact cause of mechanical disruption remains speculative, pyrolytic carbon seems to have the characteristic of fatigue fracture as weD as erosion damage. A connection between the two might exist
W. Klepetko, MD,a A. Moritz, MD,a J. Mlczoch, MD,b H. Schurawitzki," E. Domanig, MD,a and E. Wolner, MD,a Vienna. Austria
Mechanical disruption has been described in almost every type of artificial valve.!" The Edwards-Duromedies bileaflet valve (Baxter Edwards Divisions, Baxter Healthcare Corporation, Irvine, Calif.) has now been used for 4 years at our institution, and we5• 6 have reported satisfying clinical results and no instances of mechanical failure or disruption. Recently, two of our patients have experienced leaflet fracture. Complete clinical workup of these two events has been done at our institution, and both valves were investigated by the manufacturer for underlying mechanical reasons for failure. We believe these events to be the first reported cases of leaflet fracture in Edwards-Duromedies valves in the open literature.
Case reports CASE 1. A 41-year-old man underwent mitral valve replacement in October 1983 because of mixed rheumatic mitral valve failure. An Edwards-Duromedics 31 mm mitral valve
From the Second Surgical Department,' University of Vienna (Head: Prof. E. Wolner), the Cardiological Department," University of Vienna (Head: Prof. F. Kaindl), and the Department of Radiology,' University of Vienna (Head: Prof. H. Pokiesser), Vienna, Austria. Received for publication Feb. 7, 1988. Accepted for publication June 13, 1988. Address for reprints: Walter Klepetko, MD, Second Surgical Department, University of Vienna, Spitalgasse 23, A-I090 Vienna, Austria.
90
prosthesis was implanted and the postoperative course was uneventful. At follow-up of 3, 6, 12, and 24 months the patient was in New York Heart Association class I. He had no complaints and his anticoagulation program was well controlled. The last echocardiographic examination was performed in May 1986. It revealed slightly reduced left ventricular function and normal function of the prosthesis. In December 1986 the patient had sudden severe dyspnea while running up several stairs. He was then admitted to the department of cardiology because of pulmonary edema. In the acute phase the echocardiogram revealed good left ventricular function, but the function of the mitral valve prosthesis could not be determined because of tachycardia and tachypnea. After stabilization of the patient, the opening of only one leaflet could be detected; opening dysfunction of the other leaflet was, therefore, suspected. Doppler investigation revealed no significant mitral regurgitation, but the mitral flow velocity was increased to 3 msec, and the diagnosis of acute valve dysfunction was made. The patient was referred for operation immediately thereafter. At operation, the mitral valve was found to be well healed in, but one leaflet was missing; the remaining leaflet was moving freely. There were no obvious signs of tissue overgrowth or thrombus formation. The missing leaflet could not be detected within the cardiac cavity. The Edwards-Duromedies valve was excised and a Bjork-Shiley Monostrut 31 nun mitral valve (Shiley Incorporated, Irvine, Calif.) was implanted. The patient recovered quickly. On the fourth postoperative day ultrasound investigation of the great vessels revealed two foreign bodies in the iliac arteries, one on each side. This finding was confirmed by computed tomography (Fig. I). Both iliac arteries were explored by a retroperitoneal approach and two leaflet pieces were extracted. Further recovery of the patient was uneventful. Investigation of the leaflet showed a transverse fracture
Volume 97 Number 1
Edwards-Duromedics bileaflet valves
January 1989
9I
Fig. 1. Computed tomographic scan of the iliac arteries showing an embolized leaflet.
dividing it into two pieces of approximately one third and two thirds. On the excised valve, the highest part of the pyrolytic carbon housing was missing around about half the circumference on the side of the escaped leaflet. The outer metal ring, which was overgrown with a thin layer of tissue, was visible (Fig. 2). CASE 2. A 15-year-old boy was operated on because of rheumatic double valve disease in January 1984. He underwent double valve replacement with Edwards-Duromedics valves: 23 mm aortic and 31 mm mitral. Recovery was uneventful and he was discharged, but he refused any followup control studies because of lack of complaints. Thirty-six months after the operation, he suddenly had severe dyspnea during a walking tour and his condition rapidly deteriorated. On admission to the cardiological department, he was in acute pulmonary edema with severe tachypnea and tachycardia. After his condition was stabilized with medical treatment, echocardiograms were obtained. Left ventricular function was found to be good. The function of the aortic and mitral prostheses, however, could not be determined. Doppler investigation suggested light aortic regurgitation . but no mitral regurgitation could be shown. However, aortography gave no evidence of aortic insufficiency. Right heart catheterization revealed a v-wave of 80 mm Hg of pulmonary capillary wedge pressure. On fluoroscopic examination, abnormal movement of the mitral valve housing within the anulus was seen. which indicated unusual forces acting on the housing. Therefore, the patient was referred for exploration of the mitral valve. At reoperation, one leaflet of the mitral valve was missing. There was neither tissue overgrowth nor signs of thrombosis, but two small perivalvular leaks of 3 and 5 mm diameter were detected. The mitral valve was excised and a 29 mm BjorkShiley Monostrut valve was implanted . On the fifth postoperative day, ultrasound investigation of the iliac arteries was performed, but no foreign body could bedetected. By means of computed tomography, the leaflet was imaged in the left common iliac artery. It was then removed by a retroperitoneal approach. Further recovery was uneventful. Investigation of the excised valve revealed a normal valve
Fig. 2. Explanted valve from the first 'patient together with the two pieces of the escaped leaflet. Tissue ingrowth can be noted on the housing fracture surface.
housing and normal motion of the remaining leaflet. The escaped leaflet was fractured through one of the pivots (Fig. 3).
In vitro examination Both valves were returned to the manufacturer and subjected to a series of examinations in accordance with a standard protocol to determine facts about the events. The valves were photographed by macrophotographic and microphotographic techniques. X-ray films were made to examine subsurface structures, and a pathologic examination of the tissue was done. After these examinations, microscopic visual exams at 40X magnification were done to look for any unusual features on the valve. The suture rings were then removed and the valves were cleaned of all biologic debris. Scanning
92
The Journal of Thoracic and Cardiovascular Surgery
Klepetko et al.
• Fig. 3. Explanted valve from the second patient together with the major portion of the escaped leaflet. The fracture line can be seen at the upper pivot region.
Fig. 4. SEM of housing fracture surface in valve I, illustrating secondary cracking into the pivot region and the crack growth and arrest phenomenon.
electron microscopy (SEM) was performed and appropriate photographs at set magnifications were taken. Wherever feasible the valves were remeasured for compliance with original specifications and then disassembled. SEM examination was again performed after disassembly. Other nondestructive tests to determine local hardness, silicone content, or other properties were conducted. Valve 1. The pathologic examination detected tissue ingrowth over portions of the housing fracture surface. This observation supported the conclusion that this fracture had occurred long before the leaflet escaped. From light microscopic examination, there was evidence of pitting in the area of the seating surface of the housing, as well as in the pivot area of the remaining
Fig. 5. SEM of leaflet fracture surfacein valve I. The pitting can be seen on the fracture line near the edge of the leaflet.
leaflet. The escaped leaflet was in two pieces; the fracture had occurred from the seating edge at about 60 degrees from one pivot through the center of the rear of the leaflet. A pit approx imately 0.4 mm in diameter was noticeable along the fracture line. SEM examination demonstrated that the housing fracture had originated centrally and had grown circumferentially toward each pivot (Fig.A). The section fractured out at the transition from the curved seating area to the pivot flats. The fracture line was above the leaflet-housing contact line toward the outflow side of the valve. Crack growth and arrest marks in both fracture surfaces of the leaflet were demonstrated as well. The aforementioned pit at the fracture line was confirmed to blend gradually into the normal surface and abruptly into the fracture surface (Fig. 5), which suggested that the fracture originated at the pitting area. Additional smaller pits occurred on the surface of both leaflets and in the pivot ball area . Valve 2. Light microscopic studies revealed heavy chipping and pitting in the pivot slot area of the housing, both in the side of the escaped leaflet and in the opposing side. A large porous area was noted in the base of the pivot slot of the escaped leaflet opposite the side where pivot ball fracture occurred. Many porous areas around the circumference of the seating lip were also seen. Pitting was noted around the seating rim and in the pivot ball area opposite the fractured portion of the escaped leaflet (Fig. 6). The leaflet that remained in the valve had corresponding damage to its rear surface and some pitting around the seating rim, which gave evidence of interference damage. SEM examination showed a complex fracture surface through the pivot ball area (Fig. 7), the possible origin being in a slightly pitted area. The fracture surface again had evidence of
Volume 97 Number 1 January 1989
Edwards-Duromedics bileaftet valves
93
Fig. 6. SEM of housing damage at the pivot slots on the side of the pivot ball fracture in valve 2, demonstrating heavy pitting and chipping.
Fig. 7. SEM of pivot fracture surface in valve 2, showing the crack growth and arrest phenomenon.
crack growth and arrest. There was evidence of interference on the rear surfaces of the leaflets. Both pitting and mechanical damage existed on this valve, and it appeared that a combination of flow and mechanical factors were at work in this fracture, Examination of all manufacturing records revealed no abnormalities and established no trends in these failures. On remeasurement of the valves, all dimensions were found to fulfill specifications . Measurements of local hardness, silicone content, and other mechanical properties were all found to be within the normal range.
artificial valve is to establish the diagnosis and initiate appropriate treatment. The acute onset of severe symptoms in these patients is highly suggestive of mechanical dysfunction, as opposed to a more gradual onset, which might result from valve thrombosis. However, confirming this suspicion is difficult, because the patients are in unstable distress, making it technically difficult to do clinical investigations. In both of our patients, neither M-mode nor Doppler echocardiography led to conclusive results. The physical examinations, electrocardiograms, and plane thoracic x-ray films reflect only the patient's condition and not the condition of the valve, but they do have the value of excluding myocardial infarction. Only the fluoroscopic examination in one patient gave evidence of disturbance of valve function. However, it still was not possible to establish a conclusive diagnosis in either case. Therefore, the key to the survival of these two patients was immediate reoperation without an exact diagnosis. On reoperation, the valve obviously has to be replaced and the cardiac cavity explored for the missing pieces; however, it is unlikely that they will be found in the heart, because they usually embolize . This is true even in double valve replacement, as the experience with the second patient shows, in whom a 31 mm leaflet passed through a 23 mm aortic valve, without disrupting its function. Embolized leaflets have not been reported to create urgent problems; therefore, the patient's condition can be allowed to stabilize before attempts to locate and retrieve the parts are made. In our patients, we tried to find the leaflets on the fourth and fifth postoperative days. This was done by computed tomographic scan and ultrasonic imaging, which proved to be complementary
Discussion Several clinical and technological aspects of these two events are remarkable. To our knowledge this is the first report of a leaflet fracture in the Edwards-Duromedics bileaflet valve, which has been in clinical use since 1982. Mechanical disruption in other bileaflet valves has been described by others . I. 7. 8 As compared to occluder escape in single disc valves? or ball valves,'? this event in bileaflet valves normally offers a better chance for adequate treatment and patient survival, since the escape of only one leaflet still leaves some residual function of the valve. Disc escape in single disc valves has been reported to result in sudden death," but single leaflet escape in a bileaflet valve has not caused sudden death in our experience. This offers valuable time , especially in patients who are in good condition before the event. Whether this situation also pertains to patients in higher New York Heart Association classes cannot be answered by our experience. The essential factor in acute dysfunction of an
94
The Journal of Thoracic and Cardiovascular Surgery
Klepetko et al.
techniques. However, one must be aware that more than one foreign body can exist, and the search for multiple pieces should be conducted as accurately as possible. The usual location for the embolized parts is the iliac artery region, where we found the leaflet components. The small fractured pivot ball and the missing housing part were not detected and probably have embolized into smaller peripheral arteries without creating symptoms. Embolized discs of the Bjork-Shiley valve, which are larger than a leaflet of a bileaflet valve, have generally been found in the abdominal aorta.': II It is remarkable that the greatest number of reported cases of mechanical disruptions have happened in mitral valves.lv" This seems to be caused by particularly strong mechanical forces acting on these valves. From the technical examinations, the nature of the fracture in both valves seems to be a fatigue fracture, as evidenced by crack growth and arrest. In the first patient, the earlier housing fracture led to tissue ingrowth, which in our opinion may have resulted in uneven closure of the leaflet. The cause of the housing fracture remains unclear. In the operating protocol, the surgeon mentioned the removal of a calcified particle through the valve orifice after the implantation, which speculatively might have had a causal relation to the housing fracture. Another possibility for microcrack initiation could arise from the manufacturing process when the leaflets are installed or the stiffening ring is placed onto the housing. The eventual fracture of the leaflet, probably starting from one of the pitted spots, may have resulted from uneven and unusual closing forces over a long period of time. In the second valve, evidence for interference between the two leaflets was found. This was seen by unusual damage in corresponding areas of the hinge mechanism and the leaflets. Again, crack growth and arrest support fatigue as the cause of the fracture. The cause of the pitting at various locations on the carbon surface of the valve remains unclear. It could result from local mechanical forces arising from instruments or interference, from local flow conditions that could create a jetting mechanism, from local variations in material properties, and finally from local arching created by defibrillation shocks. All these theories result from in vitro tests and observations that are currently in progress at the manufacturer, and evidence suggests that each of these theories has possible importance under certain circumstances. These findings underline the importance of careful manufacturing and proved quality control for all artificial valves to avoid potential release of inadequately qualified products. It also points out the necessity of careful handling of pyrolytic carbon valves and the need for awareness on the part of the surgeon that even
unnoticed damage can result in late mechanical disruption when combined with unusually high stresses in individual patients. In addition, pyrolytic carbon seems to have qualities that are not yet fully described. For instance, fatigue is an unknown phenomenon in this material, and erosion, to our knowledge, has been mentioned in only one other published report. 13 We wish to thank the Research and Development staff of Baxter Edwards CVS Division, Irvine, California, for their assistance in conducting the examinations of the explanted prostheses. REFERENCES I. Odell JA, Durandt J, Shama DM, Vythilingum S, Spontaneous embolization of a St. Jude prosthetic mitral valve leaflet. Ann Thorac Surg 1985;39:569-72. 2. Berry BE, Sheedy PF, McGoon DC. Diagnosis and management of aortic poppet embolism. Ann Thorac Surg 1974;17:504-9. 3. Davis PK, Myers JL, Pennock JL, Thiele BL. Strut fracture and disc embolization in Bjork-Shiley mitral valve prostheses: diagnosis and management. Ann Thorac Surg 1985;40:65-8. 4. Roberts AK, Lambert CJ, Mitchel BF. Embolization of disc occluder of a Wada-Cutter mitral prosthesis with survival. Ann Thorac Surg 1976;21:361-4. 5. Moritz A, Klepetko W, Grabenwoeger F, et al. Two-years experience with the Duromedics bileaflet valve prosthesis, Texas Heart Inst J 1985;12:315-22. 6. Klepetko W, Moritz A, Khunl-Brady G, et al. Implantation of the Duromedics valve prosthesis in 400 patients. Ann Thorac Surg 1987;44:303-9. 7. Hasse J. Escaped leaflet in a St. Jude Medical mitral prosthesis. In: DeBakey ME, ed. Advances in cardiac valves: clinical perspectives. New York: Yorke Medical Books, 1983:115-23. 8. Hjelms E. Escape of a leaflet from a St. Jude Medical prosthesis in the mitral position. Thorac Cardiovasc Surg 1983;31:310-2. 9. Bonnabeau RC Jr, Lillehei CWo Mechanical "ball fracture" in Starr-Edwards prosthetic valves. J THORAC CARDlOVASC SURG 1968;56:258-64. 10. Lindblom D, Bjork YO, Semb BK. Mechanical failure of the Bjork-Shiley valve: incidence, clinical presentation, and management. J THORAC CARDIOVASC SURG 1986;92:894-907. 11. Larrieu AJ, Puglia E, Allen P. Strut fracture and disc embolization of a Bjork-Shiley mitral valve prosthesis: localization of embolized disc by computerized axial tomography. Ann Thorac Surg 1982;34:192-5. 12. Marbarger JP, Clark RE. The clinical life history of explanted prosthetic heart valves. Ann Thorac Surg 1982;34:22-33. 13. Dobrova NB, lofis NA, Kozyrkin HI, Agafonov AV, Zaretskii YV, Kevorkova RA. Analysis of causes for wear and breakdown of elements in artificial cardiac valves of the disk type. Med Tekh 1987;2:8-12.