Dvnamic Behavior of Prosthetic Aortic Tissue Vdves as Viewed by High-speed Cinematography

Transcription

1 Dvnamic Behavior of Prosthetic Aortic Tissue Vdves as Viewed by High-speed Cinematography W. Gerald Rainer, M.D., Robert A. Christopher, Ph.D., Theodore R. Sadler, Jr., M.D., and Alan D. Hilgenberg, M.D. ABSTRACT Using a valve testing apparatus of our own design and with a high-speed (600 to 800 frames per second) 16 mm movie camera, films were made of Hancock porcine, Carpentier-Edwards porcine, and Ionescu-Shiley bovine pericardial valves mounted in the aortic position and cycled under physiological conditions at 72 to 100 beats per minute. Fresh and explanted valves were observed using saline or 36.5% glycerol as the pumping solution. When fresh valves were studied using saline solution as the pumping fluid, the Hancock and Carpentier-Edwards porcine valves showed highfrequency leaflet vibration, which increased in frequency with higher cycling rates. Abnormal leaflet motion was decreased when glycerol was used as the blood analogue. The Ionescu-Shiley bovine pericardial valve did not show abnormal leaflet motion under these conditions. Conclusions drawn from tissue valve testing studies that use excessively high pulsing rates and pressures (accelerated testing) and saline or water as pumping solutions cannot be transposed to predict the fate of tissue valves in a clinical setting. The ability to predict, with reasonable confidence, the ultimate longevity and fate of artificial heart valves has met with varying degrees of success in the past. Accelerated testing and standardization of testing procedures can be applied with reasonable uniformity in evaluating valves made with metal, plastic, and other nonbiological components. However, the widespread acceptance of tissue valves in recent years has added a new dimension to the implementation of testing procedures, the in- From the Department of Research, St. Joseph Hospital, Denver, and the Department of Mechanical Engineering, University of Colorado, Boulder, CO. Supported by a grant from the Colorado Heart Association. Presented at the Fifteenth Annual Meeting of The Society of Thoracic Surgeons, Jan 15-17, 1979, Phoenix, AZ. Address reprint requests to Dr. Rainer, 2005 Franklin, Suite 700, Denver, CO terpretation of data gathered, and the transference of conclusions concerning the ultimate fate of the valve under more nearly physiological conditions. In this report we detail our experience to date in the evaluation of tissue valves by studying their behavior with high-speed cinematography and discuss possible pitfalls in transposition of certain data obtained in vitro to clinical conditions. Materials and Methods Using a valve testing apparatus of our own design (Fig 1) and previously described [lo, ll], studies were conducted on valves of the following types: Hancock porcine, Carpentier- Edwards porcine, and Ionescu-Shiley bovine pericardial. During this study, all valves were mounted in the aortic position. Flow through the system was pulsatile, and pressures and flow rates were maintained within physiological limits. Valves were cycled at 72,88, and 100 beats per minute (bpm). Using a motorized Eastman high-speed 16 mm camera with a custom-designed lens system, movies to ascertain valve behavior were made at 600 to 800 frames per second. The angle of the camera to the valve was approximately 60 degrees from the perpendicular, and the highly-polished Lucite housing caused minimal optical distortion. The system was primed with either 0.9% saline solution with added formaldehyde (10 ml of 4% solution per 20 liters of saline) or with 36.5% glycerol solution as a blood analogue. Valves tested were fresh, unused valves supplied by the manufacturers* or explanted from patients after varying survival periods. Fresh valves were removed from storage solutions for as brief a period of time as possible (usually not in excess of one hour during any *Hancock Laboratories, Anaheim, CA; Edwards Laboratories, Santa Ana, CA; Shiley Laboratories, Irvine, CA by W. Gerald Rainer

2 275 Rainer et al: Aortic Valves Viewed by High-speed Cinematography Fig 1. Valve testing apparatus showing the position of the valve mount (arrow) in the Lucite housing and the angle of the high-speed camera. test run); after the test, priming solution was rinsed off and the valve was replaced in glutaraldehyde or formaldehyde for continued storage. Prolonged exposure to glycerol was avoided [7, 161, and accelerated testing was not performed on any valve in this study. Results Analysis of films made when fresh valves and saline prime were used showed asynchrony of leaflet opening in the Hancock porcine (Fig 2) and Carpentier-Edwards porcine (Fig 3) valves. During end-systole, one or more leaflets of these valves showed full leaflet undulation and high-frequency leaflet tip vibration, always occurring in a septa1 leaflet with muscle bar where present. Frequency of undulation and flutter increased with an increasing cycle rate (up to 100 bpm in this study) and became strikingly less noticeable or absent in a glycerol prime solution or when an explanted valve was used for study. The Ionescu-Shiley bovine pericardial valve, whether fresh or explanted and whether tested in saline or glycerol solution, showed no unusual leaflet motion and leaflet opening was almost uniformly synchronous (Fig 4). Changes in cycling rate did not alter these observations. Comment The intensity of the abnormal leaflet motion seen in our studies would appear to be of the magnitude that would produce tissue damage if sustained. This might explain the discrepancy between some reported studies [2] in which major fatigue occurred under experimental conditions at 34 million to 65 million cycles (1 to 2 equivalent-years) and clinical reports [3, 5, 8, 9, 13-15] suggesting acceptable durability of tissue valves up to 7 years. Clark and colleagues [2] suggested that in vitro and clinical data for tissue valves do not correlate, and we would agree. In contrast to the high-frequency leaflet vibration noted in the Hancock porcine and Carpentier-Edwards porcine valves, there is essentially no vibration or opening asynchrony seen in the behavior of the Ionescu-Shiley bovine pericardial valve. Factors to consider in the explanation of this phenomenon include thicker leaflets in these valves and the uniformity of leaflet size and shape of the Ionescu valve compared with the natural porcine configuration. Based on our observations, the two major factors responsible for the discrepancy between the laboratory and clinical findings in tissue valves are choice of priming solution and rate of cycling. Wright [16, 171 has proposed that the priming solution of choice in tissue valve testing is saline, and he states that a mixture of glycerin and water can cause leaflets to stiffen over a period of a few minutes. This is in variance with our observations; however, we have carefully avoided prolonged exposure of tissue valves to glycerol and have concerned ourselves with valve behavior rather than emphasizing gradient study. Certainly, for the short periods of study required in our testing, we have not been impressed with any change in flexibility of leaflets in fresh, unused, carefully handled valves. On the other hand, explanted valves that we have studied do not show a significant abnormal leaflet motion even when studied in saline solution. We would suggest that the plasma coating during implantation (or in combination with fixative after explantation) may stiffen the valve leaflet enough to minimize leaflet flutter especially along the thinner free leaflet edge. The second consideration-rate of cyclingseems important from two viewpoints: increased frequency of vibration with increased cycling rate and abnormally high pressures (usually exceeding 200 mm Hg) required in

3 276 The Annals of Thoracic Surgery Vol 28 No 3 September 1979 Fresh Hancock-Aortic 23mm Saline Llmin. 80 bpm Fig 2. Sequential (every sixth frame) photographs of a fresh Hancock porcine valve. Points A and B are reference points common for each frame numbered in sequence. The line drawing is a composite design to depict leaflet flutter and vibration.

4 277 Rainer et al: Aortic Valves Viewed by High-speed Cinematography Fresh Carpentier-Edwards Aortic-23mm Saline Llmin. 72 bpm Fig 3. Sequential photographs of a fresh Carpentier- Edwards valve. See Figure 2 for details.

5 278 The Annals of Thoracic Surgery Vol 28 No 3 September 1979 Fresh lonescu-shiley Universal -27mm Saline 124/ Llmin. 72 bpm Fig 4. Sequential (every sixth frame) photographs of a fresh Ionescu-Shiley valve. No line drawing is necessary because leaflet flutter does not occur.

6 279 Rainer et al: Aortic Valves Viewed by High-speed Cinematography most accelerated testing systems. The effect of leaflet vibration has been alluded to already and higher frequency vibrations simply accelerate the process of tissue destruction. Acclerated testing may involve cycling rates up to 2,100 bpm. Greatly increased pressures are required to accomplish this rate and most likely contribute further to tissue fatigue and destruction [71. Finally, clinical studies to date have not borne out the fatigue problems as predicted by laboratory studies in which nonphysiological conditions prevail. Isolated accounts [1,4, 6, 121 of tissue valve failure have been reported but the incidence of failure due to tissue disruption of fatigue has been impressively small. Currently, it would seem appropriate to be circumspect in transference of laboratory findings to the clinical environment. Further and more sophisticated testing under physiological conditions will be worthwhile. But the ultimate fate of the valves will be disclosed by close, methodical clinical follow-up and careful examination of retrieved valves. References 1 Broom ND: Fatigue-induced damage in gluteraldehyde-preserved heart valve tissue. J Thorac Cardiovasc Surg 76:202, Clark RE, Swanson WM, Kardos JL, et al: Durability of prosthetic heart valves. Ann Thorac Surg 26:323, Cohn LH, Sanders JH Jr, Collins JJ Jr: Aortic valve replacement with the Hancock porcine xenograft. Ann Thorac Surg 22:221, Fishbein MC, Gissen SA, Collins JJ Jr, et al: Pathologic findings after cardiac valve replacement with glutaraldehyde-fixed porcine valves. Am J Cardiol 40:331, Hannah H 111, Reis RL: Current status of porcine heterograft prostheses: a 5-year appraisal. Circulation 54:Suppl 3:27, 1976 h Housman LB, Pitt WA, Mazur JH, et al: Mechanical failure (leaflet disruption) of a porcine aortic heterograft: rare cause of acute aortic insufficiency. J Thorac Cardiovasc Surg 76:212, lonescu MI: Discussion of Tandon et al lonescu MI, Tandon AP: Long-term clinical and haemodynamic evaluation of the Ionescu-Shiley pericardial xenograft heart valve. Thoraxchirurgie 26:250, Ionescu MI, Tandon AP, Mary DAS, et al: Heart valve replacement with the Ionescu-Shiley pericardial xenograft. J Thorac Cardiovasc Surg 73:31, Rainer WG, Christopher RA, Sadler TR Jr: Some comments on heart valve testing and other observations. Med Instrum 11:104, Rainer WG, Sadler TR Jr, Christopher RA, et al: High-speed cinematographic studies of clothcovered prosthetic aortic valve. Med Instrum 71283, Spray TL, Roberts WC: Structural changes in porcine xenografts used as substitute cardiac valves. Am J Cardiol 40:319, Stinson EB, Griepp RB, Oyer PE, et al: Longterm experience with porcine aortic valve xenografts. J Thorac Cardiovasc Surg 739, Tandon AP, Sengupta SM, Lukacs L, et al: Long-term clinical and hemodynamic evaluation of the Ionescu-Shiley pericardial xenograft and the Braunwald-Cutter and Bjork-Shiley prostheses in the mitral position. J Thorac Cardiovasc Surg 76:763, Tandon AP, Smith DR, Whitaker W, et al: Long-term haemodynamic evaluation of aortic pericardial xenograft. Br Ht J 40:602, Wright JTM: A pulsatile flow study comparing the Hancock porcine xenograft aortic valve prostheses models 242 and 250. Med Instrum 11:114, Wright JTM: Hydrodynamic evaluation of tissue valves, in Tissue Heart Valves. Edited by MI Ionescu. London, Butterworth, 1978, pp 13, 48, 49 Discussion DR. RICHARD EDWIN CLARK (St. Louis, MO): Last year we reported to this Society our experience with accelerated fatigue testing of a variety of valves including the tissue valves, which, much to our surprise, fatigued at 35 to 65 million cycles. This short durability in a machine was totally out of keeping with the clinical experience with porcine valves. Like Dr. Rainer, we, too, have been interested in why these valves broke so early in an accelerated fatigue tester. Consequently, during the past year we have tested a variety of tissue valves in a new, closed, lowvolume accelerated tester in which the geometry of the human aortic root is simulated in Lucite. This was to overcome one of the criticisms of our previous work. It was thought by some critics that the flow patterns had a major influence on the durability of leaflet valves in an accelerated tester. Second, we used sterile fresh-frozen human plasma matched for blood type to overcome the criticism of use of glutaraldehyde media, which was employed in previous studies. We found that the tissue valve tested broke at the same rate and fatigued significantly at the same number of cycles in the new tester as did the same valves in the old one that did not have human plasma or simulated anatomy. The cycling rate in the new tester was 20 cycles per second versus

7 280 The Annals of Thoracic Surgery Vol 28 No 3 September cycles per second used previously. Thus, the tissue valves tend to break early in a range of cyclic rates of 20 to 30 cycles per second. It is clear that the defects in these valves tested in vitro are exactly the same type of defects now appearing in clinical reports. Although we cannot predict when any particular valve is going to fail in any given patient, the sites of fatigue are now known. DR. RAINER: We are indebted to Dr. Clark and his group for their elegant studies in the past and appreciate his comments. We continue to be concerned as to the exact reason why the findings in the labo- ratory cannot be transposed to the clinical behavior of these valves. Certainly, under accelerated testing conditions, not only is the rate much faster than under physiological conditions, but the pressures required to generate these rates usually are high also. I would say that, regardless of Dr. Clark s findings and conclusions, the failure rate clinically of tissue valves certainly falls within an acceptable range and certainly is lower than nontissue valves at present. Although Dr. Clark does say that the valves will fail in the same way, it seems to me that if the numbers are acceptably small over a period of 7 to 8 years, or longer, this is an acceptable rate of failure.