Secondary Logo

Effects of Continuous Flow Left Ventricular Assist Device Support on Skin Tissue Microcirculation and Aortic Hemodynamics

Litwak, Kenneth N.*; Kihara, Shin’ichiro*; Kameneva, Marina V.*; Litwak, Philip*; Uryash, Arkady*; Wu, Zhongjun*; Griffith, Bartley P.

ORIGINAL ARTICLES
Free

Continuous flow ventricular assist devices (CFVADs) are thought to be the next generation of circulatory assist devices. With many now in various stages of development or clinical trial, it is important that the physiologic aspects of these pumps be critically analyzed. In this study, 15 calves were divided into two groups. One group received a CFVAD, and the other a sham implant. Two additional animals were used in an acute study to examine aortic blood flow patterns from a CFVAD. Tissue perfusion was measured on all animals before surgery and then weekly thereafter. Before surgery, there was no difference in hemodynamics or tissue perfusion between studied animals. Postoperatively, CFVAD animals had statistically significant increased diastolic pressure. Significantly decreased pulse pressure, pulse index, and tissue perfusion were also observed in CFVAD animals. Results from the flow pattern studies suggested that at moderate levels of pump support (40–75%), the amount of blood flow distal to the outflow graft anastomosis decreased approximately 25% because of increased regurgitant blood flow in the aorta. These results suggest that the diminished tissue perfusion is likely due to changes in aortic hemodynamics and provide some insight into the distribution of flow from CFVADs.

From the *Department of Surgery, University of Pittsburgh, Pennsylvania; and the

†Department of Surgery, University of Maryland, Baltimore, Maryland.

Submitted for consideration February 2002; accepted for publication May 2002.

Correspondence: Dr. Kenneth Litwak, University of Pittsburgh, 215 McGowan Institute, 3025 E. Carson St., Pittsburgh, PA 15203.

Continuous flow ventricular assist devices (CFVADs) are seen as the next generation of circulatory assist devices. There are many types of these devices now in stages of development or in the early stages of clinical trials. 1–5 Associated with these devices are many hemodynamic differences compared with the normal circulation and the currently available pulsatile ventricular assist devices (VADs). The primary difference is decreased pulse pressure with increased pump flow. This is due to continuous pump flow during the diastolic portion of the cardiac cycle, raising diastolic blood pressure. 3 Another major difference is the potential for regurgitant pump flow during periods of hypertension or pump stoppage, as these pumps have no valves. 6 Moreover, the distribution of pump flow in the aorta and the potential effects of such changes on overall blood flow are poorly documented. 6–8 Changes in aortic hemodynamics could have profound implications for end-organ function, as well as vascular structure and function. Indeed, there have been recent reports of changes in aortic structure and function with continuous flow pump support. 8 Discussions of microvascular blood flow with CFVAD support are equally limited, mostly reserved for coronary microcirculation or nondescript end-organ function. 9–13 The purpose of this study was to examine the effects of chronic continuous flow pump support on skin tissue microcirculation. Furthermore, in an effort to relate these effects to blood flow patterns, we conducted acute studies to examine aortic hemodynamics with varying amounts of continuous flow pump support.

Back to Top | Article Outline

Materials and Methods

Chronic Studies: Implantation Procedure

Fifteen male Jersey breed calves (55–70 kg) were divided into two groups: pump implant (PI, n = 8) and sham implant (SI, n = 7). In the PI group, continuous flow pumps were implanted in the left thoracic cavity in a left ventricular apex to descending thoracic aorta bypass. Briefly, animals were preanesthetized with atropine sulfate (30–45 mg subcutaneously) and then anesthetized with Methohexital (10 mg/kg intravenously). Animals were immediately intubated and anesthesia was maintained with isoflurane (1.5–3.0%) mixed with oxygen and room air. The carotid artery was permanently catheterized with a polypropylene catheter. A 5th rib interspace thoracotomy was done. The dacron outflow graft was preclotted in the animal’s blood and then anastomosed onto the descending thoracic aorta. The pump was assembled and the apical cannula was inserted directly into the left ventricular apex, after the apex had been cored with a custom circular coring knife. The pump circuit was de-aired with multiple needle punctures in the outflow graft. Main pulmonary artery (PA) and pump flow were measured using implanted ultrasonic flow meters (Transonic Animal Research Flowmeters T206 series, Transonic System Inc., Ithaca, NY). Pump speed was adjusted so that pump flow was 50–80% of PA flow. The thoracotomy was closed in a three layer pattern. Anticoagulant therapy (coumadin, 2.5–12.5 mg orally) was administered the night before the surgery and then daily thereafter. International normalized ratios (INRs) were adjusted twice weekly to maintain an INR of 2.5–3.5.

SI animals were also subjected to a thoracotomy. A short (2–4 cm) length of preclotted dacron graft was anastomosed onto the descending thoracic aorta, and the end was sewed over. A small incision was made in the left ventricular apex, and a custom circular coring knife was used to remove a small amount of the ventricular tissue; the LV defect was then closed. Thoracotomy closure was identical to implanted animals. SI animals were anticoagulated in the same manner as PI animals.

Blood collection was carried out on the same schedule for both groups. Hematocrit and hemoglobin were measured twice weekly. Blood pressure measures for analyses were measured at the same time as tissue perfusion.

Back to Top | Article Outline

Skin Tissue Perfusion

Skin tissue perfusion was measured before surgery, during surgery, and then weekly after surgery with a laser doppler flow meter (Model BLF21D, Transonic). 14 The probe was secured to the ventrum of the tail at its juncture to the body. This location was chosen as it provided a flat area, with thin skin, where the probe could be reproducibly secured. As such, motion artifact was minimized. Furthermore, this location allowed for measurement with minimal animal restraint. Measurements were made over a 1 minute period while the animal was standing quietly. Tissue perfusion at each time point was determined as the average of each measurement period. In two pump implanted animals, tissue perfusion was measured as pump flow was changed. These two animals were awake and standing quietly during the trial. Tissue perfusion was measured at five different pump flow readings, ranging from 20 to 85% of PA flow.

Back to Top | Article Outline

Flow Study

Two calves had acute procedures done to determine aortic hemodynamics with continuous flow pump support. Anesthesia was done as previously described. The thoracotomy was done through the bed of the 5th rib, which was resected. A continuous flow pump was placed in a left ventricle to descending thoracic aorta bypass. Flow probes were placed on the PA and pump outflow graft. An additional two flow probes were placed on the aorta, 3 cm to each side of the outflow graft anastomosis (referred to in the text as proximal and distal aorta). Measurements were made as follows: (1) pump circuit clamped (baseline); (2) pump circuit open, pump off; (3) mean pump flow = 25–100% of mean PA flow; and (4) heart failure induced with esmolol (250 mg iv) with mean pump flow = 25–100% of mean PA flow. Heart failure was quantified as PA flow less than 50% of baseline. Data points were collected every 5 ms for 30 seconds. Data for pump flow and aortic flow were normalized as a percent of PA flow to remove bias due to differences in PA flow rates.

Back to Top | Article Outline

Statistical Analysis

All data are expressed as mean ± standard error. Average systolic, diastolic, and mean blood pressure measurements were derived for each animal from weekly measurements, after which a group average was determined. The paired t-test and repeated measures ANOVA were used to test statistical significance, with p < 0.05 considered statistically significant. All animals received humane care in accordance with the Guide for Care and Use of Laboratory Animals (NIH publication 86–23, revised 1996) and the guidelines determined by the Institutional Animal Care and Use Committee of the University of Pittsburgh.

Back to Top | Article Outline

Results

At baseline, there were no differences between PI and SI animals. After implantation, there was a significant increase in diastolic blood pressure, with a concurrent decrease in pulse pressure and pulse index in PI animals. There were no significant differences in hematocrit or hemoglobin between groups at any measured time point. PI animals had significantly decreased hematocrit and hemoglobin for the first 4 weeks postoperatively, compared with baseline (p < 0.05); however, by the termination of the study, these values had increased such that there was no difference from baseline. SI animals showed a similar trend, although there were no significant differences from baseline at any measured time point.

Back to Top | Article Outline

Tissue Perfusion

There was no difference in tissue perfusion at baseline between PI and SI animals (12.4 ± 1.1 vs. 13.9 ± 1.4, p = 0.43). However, 1 week after implant and continuing out for the 5 weeks measured, PI animals had a significant decrease in tissue perfusion from baseline and when compared with SI animals (Figure 1). There was no correlation between this decrease and hematocrit or hemoglobin concentrations at any time point. Furthermore, when pump flow was decreased, there was a linear decrease in tissue perfusion (Figure 2) (r = 0.93, p < 0.01) below that which it had already dropped postimplant.

Figure 1

Figure 1

Figure 2

Figure 2

Back to Top | Article Outline

Aortic Hemodynamic Studies

Data from the aortic hemodynamic studies are presented in Figures 3 and 4. In the normal bovine aorta, aortic flow at the location measured represents approximately 40% of the cardiac output. There is a large surge of blood through the aorta during systole, which gradually tapers off towards zero normograde flow (with occasional momentary retrograde flow) during diastole. The time over which there is zero or retrograde flow in the aorta is less than 5% of the cardiac cycle.

Figure 3

Figure 3

Figure 4

Figure 4

At pump flow of 50% cardiac output, changes to the normal blood flow were as follows: (1) in the aorta proximal to the outflow graft (proximal aorta), there was a spike of normograde blood flow from the heart, followed by a spike of retrograde blood flow from the pump, followed again by a small peak of normograde blood flow; and (2) in the aorta distal to the outflow graft (distal aorta), there was a larger spike of normograde blood flow corresponding to positive flow from the heart and pump, followed by a long time period of zero or retrograde flow corresponding to the net zero flow through the pump.

At pump flow of 100% of cardiac output, changes to the normal blood flow were as follows: (1) in the proximal aorta, flow patterns were similar to those demonstrated at pump flow of 50% cardiac output, but with less pulsatility to the flow; and (2) in the distal aorta, there was virtually no pulsatility to the flow wave.

When the pump was off, with the circuit open (mimicking pump failure), changes to the normal blood flow were as follows: (1) in the proximal aorta, the flow pattern was generally similar to a normal pattern, although with a greater spike; and (2) in the distal aorta, there was a similar spike of normograde flow corresponding to systole; however, during diastole, all normograde flow from the proximal aorta was directed retrograde through the outflow graft, with the net effect of zero normograde blood flow in the distal aorta for approximately 50% of the cardiac cycle.

To further quantify the effects of different levels of pump support on distal aorta blood flow, pump flow and distal aorta blood flow were normalized to PA blood flow (Figure 5). The results suggest that at pump flows of 25% and 100% of PA flow, there is a normal amount of blood flow through the distal aorta. However, at intermediate pump flows, there is a trend towards a decrease in blood flow through the distal aorta (pump flow = 25% PA flow vs. Pump flow = 50% PA flow, p = 0.08). During conditions of induced heart failure, blood flow through the distal aorta was approximately twice normal at all levels of pump support (65–75%vs. 40% of PA flow).

Figure 5

Figure 5

Back to Top | Article Outline

Discussion

Continuous flow blood pumps are seen as the next generation of VADs because of their small size, relative simplicity, lack of valves, and minimal moving parts. However, since these pumps have no valves, they operate in synchronicity with left ventricular pressure and contraction. 6 As the blood pressure in the aorta increases, there will be a corresponding decrease in pump flow at constant pump speed. Thus there is a pulsatile characteristic to the flow. Whereas this may not be very relevant in a diseased heart, which cannot generate a significant pulse, in those hearts that can or eventually regain the ability to contract, this change in pump flow characteristics could have profound implications to VAD support.

One of the most startling findings of this study was the considerable decrease in skin tissue perfusion with continuous flow pump support. As this measurement is directly related to the amount of blood flow through the skin capillaries, it appears as if there is lower capillary blood flow with continuous flow pump support. The direct correlation between pump flow rate and tissue perfusion (Figure 2) suggests that this is not due to the loss of pulsatility but a hemodynamic mismatch. Although the pulse pressure is increasing, there is a concurrent decrease in tissue perfusion. In fact, based upon the flow study data, decreasing pump flow from 100% of cardiac output to 50% (support levels similar to that of the PI animals) (Table 1), resulting in an increase in pulse pressure from 24 to 40 mm Hg, actually diminished flow in the aorta distal to the outflow graft by 25%, suggesting a partial mechanism for the 50% decrease in tissue perfusion. Decreased total blood flow is suggested by Figure 4, wherein the total time of regurgitant flow in the cardiac cycle increases from less than 5% under normal flow circumstances to 25% when pump flow is 50% of PA flow. Whereas the volume of retrograde blood flow is small, the energy equivalent pressure may be decreased, as suggested by others. 15–17 A decrease in the energy of the blood flow could decrease the delivery of blood flow to the high resistance capillary beds.

Table 1

Table 1

As pump support increased from 25 to 40% of cardiac output, there was a 25% decrease in distal aortic blood flow. This decreased level persisted until the pump support was approximately 100%. This is likely to be due to interaction of pump flow with native heart flow. At lower pump support, the heart is contributing blood flow down the aorta, and the pump is unable to overcome the pressure head from the blood leaving the heart. Thus the blood from the pump tends to be directed distally. As pump support increases, the trend is for blood flow from the pump to become more evenly divided, going both normograde and retrograde, as the pump is able to overcome the pressure head from the heart. This change in aortic blood flow could partially explain the diminished tissue perfusion seen in the PI animals and shows the role of blood pressure in pump flow and aortic hemodynamics.

Alterations in aortic blood flow could potentially explain the findings of other groups. Nishimura et al.8 describe changes to aortic structure and function after exposure to continuous flow pump support. In their study, there was a seemingly paradoxical increase in elastin and a decrease in collagen content in the aorta of goats subjected to continuous flow pump support compared with controls. Based upon our aortic flow study data, in which there is increased pulsatility of the flow wave in the aorta between the heart and the outflow graft, one could predict an increase in elastin content in that section of the aorta. Likewise, distal portions of the aorta, which are subjected to decreased flow pulsatility, may have increased collagen content, a subject of ongoing study.

Even more interesting were the distal aorta blood flow results during mock pump failure. In approximately 50% of the cardiac cycle, there was no normograde blood flow in the distal aorta. During end systole and early diastole, the majority of blood coming from the heart down the aorta goes back through the pump and does not continue down the aorta. Whereas the total volume loss is fairly low, the length of time involved and the subsequent decrease in flow wave energetics are likely significant. This suggests that should the pump fail after being implanted onto the ascending aorta, there could be a substantial drop in blood delivered to the brain, a potentially debilitating scenario.

In the heart failure model, there was an increase in blood flow through the distal aorta with the heart failure model (approximately 2× normal), exactly the opposite of what was shown in the healthy heart aortic hemodynamic studies. Whereas the cause of this is unclear, our hypothesis is that as the blood pressure decreases with the induced heart failure, there is decreased pressure head for the pump to work against in both phases of the cardiac cycle, allowing for less competition with the native heart ejection. The heart producing a small ejection may be enough to preferentially direct blood distally because of a slightly greater pressure head proximal to the outflow graft. Again, the implications are likely not as great for humans when the graft is being attached to the ascending aorta, but it could be significant for a descending thoracic aorta anastomosis.

Back to Top | Article Outline

Study Limitations

These data suggest marked effects from implantation with continuous flow blood pumps; however, they should be taken in perspective. Virtually all studies were done with animals with a healthy heart. As such, the relevance to the typical human patient with a diseased heart may be limited. If there is cardiac recovery, these data would certainly become more relevant. Furthermore, the attachment of the outflow graft was on the descending thoracic aorta. Whereas the majority of VADs are likely to be attached to the ascending aorta, the descending thoracic aorta may receive more attention as the drive for less invasive surgery increases. In addition, the data suggesting the implications of pump failure are relevant no matter where the outflow graft is located and may have more profound clinical implications if the anastomosis site is the ascending aorta (retrograde blood flow through the outflow graft may decrease overall blood flow to the brain, leading to increased likelihood of regional ischemia). Further studies are being planned to help elucidate these issues.

Back to Top | Article Outline

Acknowledgments

This work was supported in part by a grant from the National Institutes of Health, 1R01 HL64950–01. The authors thank Mary Watach and Lisa Gordon for their excellent technical assistance.

Back to Top | Article Outline

References

1. Fossum TW, Morley D, Benkowski R, et al: Chronic survival of calves implanted with the DeBakey ventricular assist device. Artif Organs 23: 802–806, 1999.
2. Olsen DB: The history of continuous-flow blood pumps. Artif Organs 24: 401, 2000.
3. Frazier OH, Myers TJ, Jarvik RK, et al: Research and development of an implantable, axial-flow left ventricular assist device: The Jarvik 2000 Heart. Ann Thorac Surg 71: S125–132, 2001.
4. Clark RE, Walters RA, Hughson S, Davis SA, MaGovern GJ: Left ventricular support with the implantable AB-180 centrifugal pump in sheep with acute myocardial infarction. ASAIO J 44: 804–811, 1998.
5. Burke DJ, Burke E, Parsaie F, et al: The HeartMate II: Design and development of a fully sealed axial flow left ventricular assist system. Artif Organs 25: 380–385, 2001.
6. Akimoto T, Yamazaki K, Litwak P, et al: Relationship of blood pressure and pump flow in an implantable centrifugal blood pump during hypertension. ASAIO J 46: 596–599, 2000.
7. Nishimura T, Tatsumi E, Takaichi S, et al: Morphologic changes of the aortic wall due to reduced systemic pulse pressure in prolonged non pulsatile left heart bypass. ASAIO J 43: M691–695, 1997.
8. Nishimura T, Tatsumi E, Taenaka Y, et al: Effects of long-term nonpulsatile left heart bypass on the mechanical properties of the aortic wall. ASAIO J 45: 455–459, 1999.
9. Reddy RC, Goldstein AH, Pacella JJ, Cattivera GR, Clark RE, Magovern GJ Sr: End organ function with prolonged nonpulsatile circulatory support. ASAIO J 41: M547–M551, 1995.
10. Hata M, Shiono M, Orime Y, et al: Coronary microcirculation during left heart bypass with a centrifugal pump. Artif Organs 20: 678–680, 1996.
11. Lee J, Menkis AH, Tyml K, et al: Microvascular comparison between pulsatile and non-pulsatile perfusion, in Schima H, Thoma H, Wiesethaler G, Wolner E (eds), Proceedings of the International Workshop on Rotary Blood Pumps. Vienna, Italy, 1991, pp. 177–183.
12. Golding LR, Murakami G, Harasaki H, et al: Chronic nonpulsatile blood flow. Trans Am Soc Artif Intern Organs 28: 81–85, 1982.
13. Yada I, Golding LR, Harasaki H, et al: Physiopathological studies of nonpulsatile blood flow in chronic models. Trans Am Soc Artif Intern Organs 29: 520–525, 1983.
14. Hales JR, Stephens FR, Fawcett AA, et al: Observations on a new non-invasive monitor of skin blood flow. Clin Exp Pharmacol Physiol 16: 403–415, 1989.
15. Undar A, Frazier OH, Fraser CD Jr: Defining pulsatile perfusion: Quantification in terms of energy equivalent pressure. Artif Organs 23: 712–716, 1999.
16. Pantalos GM, Marks JD, Riebman JB, Burton NA, DePaulis R, Kolff WJ: Hemodynamic and energetic assessment of calves implanted with a left ventricular assist device (LVAD). Int J Artif Organs 11: 119–126, 1988.
17. Wright G: Hemodynamic analysis could resolve the pulsatile blood flow controversy. Ann Thorac Surg 58: 1199–1204, 1994.
Copyright © 2003 by the American Society for Artificial Internal Organs