Cavitation is the formation of vaporous (or gaseous) cavities (bubbles) caused by transient reduction in local pressure below liquid vapor pressure and their subsequent collapse during pressure recovery. In the case of mechanical heart valves, cavitation is produced by pressure variations in the blood associated with valve geometry and dynamics of valve closure.1 The short duration of the mechanical heart valve cavitation of less than 1 millisecond suggests that it is of vaporous type.
Upon bubble collapse, high forces, fluid jets, and pressure and thermal shock waves are produced that can cause blood and valve surface damage.2 The cavitation erosion of the valve components has been reported in early in vitro studies as well as the surface pitting on valves explanted from animals and humans.3–5 Additionally, blood damage due to cavitation has been observed in vitro6–9 and in vivo.10 Several studies have shown that cavitation associated with valves in pulsatile assist devices, when present, was the predominant cause of hemolysis.7–9 Therefore, the minimization of cavitation when designing pulsatile assist devices is of high importance.
In the past, cavitation was studied visually in transparent blood analog fluids using stroboscopic photography and videography.11,12 The cavitation cycle time and the area where cavitation was present were used to quantify cavitation intensity. Garrison et al.7 showed that high-frequency pressure oscillations associated with the bubble formation and collapse correlate with the visual appearance of the cavitation bubbles. The high-fidelity pressure transducer (hydrophone) mounted in the proximity of the valve was used to detect these oscillations. This method allowed for measuring cavitation intensity in transparent and opaque fluids such as blood.
Cavitation has been studied with respect to valve geometry, occluder material, and occluder-housing gap widths.13–15 The maximum and mean ventricular dP/dt were initially used to compare the cavitation threshold.1 Subsequent development of the laser-sweeping technique allowed for comparisons based on the valve closing dynamics.16 It was shown that elevated valve closing velocity and deceleration, which are related to the elevated dP/dt, play a role in cavitation inception.15,17 Increased cavitation intensity, with decreased ventricular filling, was also shown7–9 and is thought to be related to the increased valve closing velocity.15
A pulsatile, pneumatically actuated, pediatric ventricular assist device (VAD) is currently under development at Penn State University. This device is similar in design to the adult-size Pierce-Donachy VAD (Thoratec VAD).18 The dynamic stroke volume ranges from approximately 10 to 15 ml, depending on valve type and operating conditions. Two valve designs are currently under consideration: the Björk Shiley Monostrut (Alliance Medical, Irvine, CA) and the CarboMedics bileaflet valves (CarboMedics, Austin, TX). In this study, cavitation associated with the Björk Shiley valves was investigated. Particularly, the effect of diastolic duration on inlet valve cavitation was investigated using a high-fidelity hydrophone. Although, the risk of cavitation is higher for the valve in the mitral position,13 very fast closure of the outlet valve (determined by the vacuum in the driver unit) can cause conditions prone to cavitation. Therefore, the effect of systolic duration on the outlet valve cavitation was also studied.
Materials and Methods
Mock Circulatory Loop Design
A new, low-volume (280 ml) mock circulatory loop was assembled for the purpose of this and subsequent hemolysis studies (Figure 1). The adult-size Pierce-Donachy pneumatic pump without valves was used as an aortic compliance whereas a venous reservoir was used as a venous compliance. To provide the adequate aortic compliance, the air side of the pneumatic pump was connected to a 0.3-l glass flask in which the mean air pressure was regulated to approximately 70 mm Hg. A screw clamp positioned between the aortic and the venous compliance chambers provided systemic resistance. The mean arterial pressure was maintained at approximately 75 mm Hg with diastolic and systolic pressures of 60 to 90 mm Hg, respectively (a pediatric physiologic arterial pressure).19 Clinical-length cannulae were used to simulate the clinical pressure-flow relationships in the valve region. Because the connectors of the adult size aortic compliance were larger than the diameter of the outlet clinical cannula, the screw clamp was used to provide adequate cannula resistance. Blood-contacting surfaces in the loop were segmented poly(ether polyurethane urea), Tygon, and polished titanium tubing connectors and the working fluid was phosphate-buffered saline. The loop was instrumented to measure the following: VAD inlet pressure, VAD outlet pressure, VAD arterial pressure, VAD outlet cannula flow, VAD inlet cannula flow, VAD pneumatic driveline pressure, the inlet valve cavitation, and the accelerometer signal.
Data Acquisition Settings
Drive line, inlet, outlet, and arterial pressures were monitored throughout the experiments with Maxxim disposable transducers model 041500503A (Maxxim Medical, Inc., Athens, TX) via fluid-filled catheters. The zero pressure reference point was the center of the VAD case. The inlet and outlet flow rates were measured with the use of Transonic ultrasound flow sensors models 8C and 4XL (Transonic System Inc., Ithaca, NY) attached to Transonic ultrasonic flow meters models T206 and TS 420. The pressure transducers and flow meters were connected to the National Instruments Data Acquisition Hardware unit (SCXI-1000; National Instruments, Austin, TX). Pressure, flow, and accelerometer signals were acquired with BioBench software (National Instruments) with a sampling rate of 1,000 samples/s.
A high-fidelity hydrophone (model W132A31; PCB Piezoelectronics, Depew, NY) was mounted to the connector approximately 0.5 cm upstream from the mitral valve to measure high-frequency pressure oscillations produced by cavitation bubble formation and collapse. The signal was acquired at a 5-MHz sample rate with a digital oscilloscope (Model TDS 420, Tektronix Inc., Beaverton, OR) and then transferred to a PC computer via a National Instruments GPIB-USB-A card. A LabVIEW program (National Instruments) was written to import a 6-millisecond-long pressure signal (30,000 data points) from the GPIB card and the oscilloscope, and subsequently bandpass filtered the signal between 50 and 500 kHz with a 10th-order Butterworth filter. This filtering removed the low-frequency components of the “ringing” of the valve and the high-frequency components attributed to the resonant frequency of the pressure transducer. The remaining pressure oscillations were assumed to be those associated with the cavity formation and collapse. The RMS intensity of a 6-millisecond window pressure trace, bandpass filtered from 50 to 500 kHz, was used to quantify cavitation intensity.
To study the effect of diastolic duration on the inlet valve cavitation, the VAD rate was varied for a constant inlet pressure, outlet pressure, driving pressures, and systolic duration (300 milliseconds). The inlet valve cavitation was investigated with respect to different filling conditions altered by changing diastolic drive pressure and the inlet cannula resistance to simulate the effect of partial cannula obstruction. The effect of systolic duration on the outlet valve cavitation was examined at a fixed beat rate (86 bpm) by changing systolic duration from 250 to 330 milliseconds while keeping all the pressures constant. Throughout the entire experiment, the arterial pressure was maintained at 90/60 mm Hg with a mean of 75 ± 3 mm Hg.
Approximately 80 beats were recorded at every test condition. High-speed video and an accelerometer were used to determine the position of the valves during closure.
Results and Discussion
Figure 2a shows the driveline pressure, inlet and outlet flows, and the accelerometer signals at the condition when the onset of systole (determined by the increase in driveline pressure) occurred after the pump was completely full. The high-speed video showed that the inlet valve was almost closed at that moment and therefore the impact of the valve was gentler. The spike in the accelerometer signal signifies the valve closure (as confirmed from the video data). An example of the low-intensity cavitation signal associated with the “late” systole is shown in Figure 2c (RMS intensity was 1.5 mm Hg).
Figure 2b shows the driveline pressure, the inlet and outlet flows, and the accelerometer signals at the condition when the onset of systole occurred before the pump was completely filled. An example of the high-intensity cavitation signal typically associated with the “early” systole is shown in Figure 2d (RMS intensity was 17.3 mm Hg).
Similarly, the “late” onset of diastole (Figure 2e) caused the gentler outlet valve closure and decreased risk for cavitation, whereas the “early” onset of diastole caused faster outlet valve closure and the condition was therefore prone to cavitation (Figure 2f). The outlet valve cavitation intensity was measured with the hydrophone placed on the connector downstream of the valve, which is the opposite side of the expected cavitation occurrences. The low outlet valve cavitation intensities for all conditions studied can be attributed to the placement of the hydrophone. Examples of the cavitation signals typically associated with the “late” and “early” diastole are shown in Figures 2g and 2h.
Figure 3a shows the inlet valve cavitation intensity with respect to diastolic duration. The systolic valve duration was kept constant (300 milliseconds) while the diastolic duration was altered by changing the beat rate for a fixed condition of the inlet, arterial, and driving pressures. The five data series represent the five different filling conditions that were achieved by changing the diastolic drive pressure and, in addition, for the data series B and E, by increasing the inlet cannula resistance. For all data series, shorter diastolic time means less time for pump filling and higher cavitation intensity. Figure 3b shows the peak inlet regurgitant flow associated with the cavitation intensities shown on Figure 3a.
The outlet valve cavitation intensity with respect to the systolic duration is shown on Figure 3c. The beat rate was kept constant at 86 bpm, whereas the systolic duration was changed from 250 milliseconds to 330 milliseconds. As discussed earlier, when the diastolic drive pressure is applied and the ejection phase is not complete, the outlet valve rapidly closes, which can cause conditions prone to cavitation. Although very small, the outlet cavitation intensities have a similar decreasing trend with an increase in systolic duration as the inlet valve cavitation increases with an increase in diastolic duration. Figure 3d shows the peak outlet regurgitant flow associated with the cavitation intensities shown on Figure 3c.
Figures 4a and 4b show that both inlet and outlet valve cavitation increase with increased peak regurgitant flow. This correlation between inlet valve cavitation intensity and inlet peak regurgitant flow was noticed when running the Penn State adult-size electrical left VAD (LionHeart).9 Therefore, this parameter can be used to predict cavitation intensity of the pump under controlled experimental conditions.
An understanding of the relationship of the inlet and outlet valve cavitation to the diastolic and systolic duration can be used to determine the optimal operating conditions of the pulsatile pediatric pump and to aid in the design of the pneumatic drive system controller. To avoid conditions prone to cavitation, the systolic and diastolic durations should be adjusted so that the onset of systole and diastole occur approximately 10–15 milliseconds after the pump completely fills and empties. Prolonged time should also be avoided because extra time can cause more aggressive valve rebound and additional cavitation incidents.
Future studies should be conducted to investigate the effect of dP/dt on inlet valve cavitation and hemolysis to determine the optimal drive line compliance. The previous studies have shown increased hemolysis with increased drive line dP/dt and insufficient compliance.20
This research was funded by the National Heart, Lung, and Blood Institute contract N01-HV-48191.
1. Hwang NHC: Cavitation potential of pyrolytic carbon heart valve prostheses: A review and current status. J Heart Valve Dis
7: 140–150, 1998.
2. Young F: Cavitation
. New York, McGraw-Hill, 1989.
3. Bokros JC: Carbon in prosthetic heart-valves. Ann Thorac Surg
48: S49–S50, 1989.
4. Klepetko W, Moritz A: Leaflet fracture in Edwards Duromedics bileaflet valves. Thorac Cardiovasc Surg
97: 90–94, 1989.
5. Kafesian R, Howanec M, Ward GD, et al
: Cavitation damage of pyrolytic carbon in mechanical heart valves. J Heart Valve Dis
3: 52–57, 1994.
6. Freed D, Walker W, Dube C: Cavitation effects near moving prosthetic surfaces. Proc Am Soc Mech Eng Biomech Symp
7. Garrison LA, Lamson TC: An in-vitro
investigation of prosthetic heart valve cavitation in blood. J Heart Valve Dis
3: S8–S24, 1994
8. Lamson T, Rosenberg G, Geselowitz D: Relative blood damage in the three phases of a prosthetic heart valve flow cycle. ASAIO J
39: M626–M633, 1993.
9. Lukic B: In vitro
study of hemolysis and operating conditions in a left ventricular assist device [MS thesis]. Hershey, The Pennsylvania State University
10. Zapanta CM, Stinebring DR, Sneckenberger DS, et al
: In vivo observation of cavitation on prosthetic heart valves. ASAIO J
42: M550–M555, 1996.
11. Stinebring DR, Farrell KJ, Billet ML: The structure of a 3-dimensional tip vortex at high Reynolds-numbers. J Fluids Eng
113: 496–503, 1991.
12. Zapanta CM, Liszka EG, Lamson TC, et al
: A method for real-time in-vitro observation of cavitation on prosthetic heart-valves. J Biomech Eng
116: 460–468, 1994.
13. Graf T, Fischer H, Reul H, Rau G: Cavitation potential of mechanical heart-valve prostheses. Int J Artif Organs
14: 169–174, 1991.
14. Sneckenberger DS, Stinebring DR, Deutsch S, et al
: Mitral heart valve cavitation in an artificial heart environment. J Heart Valve Dis
5: 216–227, 1996.
15. Zapanta CM, Stinebring DR, Deutsch S, et al
: A comparison of the cavitation potential of prosthetic heart valves based on valve closing dynamics. J Heart Valve Dis
7: 655–667, 1998.
16. Guo G, Xu C, Hwang NHC: Laser assessment of leaflet closing motion in prosthetic heart valves. Biomech Eng
12: 474–481, 1990.
17. Chandran KB, Lee CS, Chen LD: Pressure field in the vicinity of mechanical valve occluders at the instant of valve closure: Correlation with cavitation initiation. J Heart Valve Dis
3: S65–S76, 1994.
18. Pierce WS, Rosenberg G, Donachy JH, et al
: Postoperative cardiac support with a pulsatile assist pump: Techniques and results. Artif Organs
11: 247–251, 1987.
19. Gregory GA: Pediatric Anesthesia
, vol 1. New York, Churchill Livingstone, 1983.
20. Ducko C, McGregor ML, Rosenberg G, Pierce WS: The effect of valve type and drive line dP/dt on hemolysis in the pneumatic ventricular assist device. Artif Organs
18: 454–460, 1994.