| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2002;123:63-71
© 2002 The American Association for Thoracic Surgery
Cardiopulmonary Support and Physiology |
From the Departments of Anesthesia,a Cardiovascular Surgery,b and Community Health Sciences,c University of Manitoba, Winnipeg, Manitoba, Canada.
The Crocus Investment Fund and the Industrial Research Assistance Program provided funding. Some of the concepts discussed are protected by US Patents #5,647,350, #5,941,841 and #6,027,498; "Control of Life Support Systems," owned by Biovar Life Support Inc, jointly held by Drs Mutch, Lefevre, the University of Manitoba, and the Crocus Investment Fund.
Received for publication April 23, 2001. Revisions requested June 7, 2001; revisions received June 21, 2001. Accepted for publication June 26, 2001. Address for reprints: W. A. C. Mutch, MD, Professor, Department of Anesthesia, University of Manitoba, A504 Chown Building, c/o 170 Services Building, 744 Bannatyne Ave, Winnipeg, Manitoba, Canada R3C 0W3 (E-mail: amutch{at}ms umanitoba.ca).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
1-3 With maximal coronary vasodilation induced pharmacologically, 4-fold differences in blood flow between regions can be demonstrated, and flow variability increases at lower perfusion pressures. 4
The distribution of cardioplegic solution is also heterogeneous, even in the absence of coronary disease, and Aldea and colleagues 5 have emphasized the importance of coronary perfusion pressure on delivery of cardioplegic solution. Standard cardioplegia is delivered with a roller pump with an apulsatile flow pattern. This may not optimally recruit all vascular territories during ischemic arrest, with the attendant increase in vascular resistance and vessel closure. Thus, there is a high likelihood that some regions of the myocardium may receive inadequate protection with standard cardioplegia.
Bassingthwaighte and colleagues 6,7 have suggested that the branching pattern of the coronary circulation is fractal in structure, which may explain the heterogeneous flow patterns observed. Fractals are structures or processes that demonstrate self-similarity with increasing magnification. Global and regional pressure flow from the myocardium can be mathematically modeled as simple power-law functions with differing exponents. 5,8,9 Such power-law functions also indicate fractal behavior.
We have developed biologically variable (BV) life support, which restores the normal variability seen in physiologic signals to either roller pumps or mechanical ventilators by using computer controllers. 10-15 BV life support may be characterized as a fractal process because the pressure-time relationships generated by either of the computer controllers also fit power-law functions.
In the vascular system BV CPB results in improved cerebral oxygenation during rewarming from moderate hypothermia compared with either apulsatile 11 or monotonously regular pulsatile CPB. 14 Mathematic modeling suggests that biologic variability, using a fractal sequence for the driving pressure, will augment recruitment of either alveoli in the respiratory system 16 or microvessels in the vascular system 14 without an overall increase in driving pressure.
We hypothesized that infusing cardioplegic solution in a fractal manner may enhance delivery, resulting in improved myocardial protection during ischemic arrest.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Surgical preparation
Twenty healthy pigs (33-38 kg) were premedicated (atropine, 0.6 mg; ketamine, 12 mg/kg; and midazolam, 0.5 mg/kg) and anesthetized with isoflurane 1.4% in oxygen, with doxacurium (8 mg/h) used for muscle relaxation. Volume-cycled ventilation through an endotracheal tube maintained arterial carbon dioxide and pH within the normal range. Electrocardiograms were recorded with surface electrodes.
Femoral arterial, pulmonary arterial, and urinary catheters were placed to monitor mean arterial blood pressure, pulmonary artery pressure, and urine volume. Through the femoral vein, an 8F balloon-tipped Fogarty catheter was passed into the inferior vena cava above the level of the diaphragm. Nasopharyngeal and left ventricular myocardial free-wall temperature probes were positioned for continuous temperature monitoring.
A sternotomy was performed, and the pericardium was opened. Left ventricular dimension was measured in 2 planes (anterior-posterior and apex-base) by pairs of 2-mm piezo-electric ultrasonic transducers positioned on the endocardial surface through a stab wound and secured with a purse-string suture. A Millar Mikro-tip 3F catheter (Millar Instruments, Inc, Houston, Tex) was secured in the left ventricle through the apex for continuous measurement of left ventricular pressure (LVP). The crystal pairs were attached to a digital sonomicrometer (Sonometrics Corporation, London, Ontario, Canada).
CPB cannulas were placed in the ascending aorta (6.5 mm; Stöckert GmbH, Freiburg, Germany) and right atrium (32F DLP single-stage cannula; DLP, Inc, Walker, Mich). A 9F cannula (Medtronic, Inc, Minneapolis, Minn) with a side branch for pressure monitoring was inserted into the aortic root for administration of cardioplegic solution. Before the start of CPB, 100 mg of sodium hydrocortisol succinate was given intravenously. The bypass circuit was primed with lactated Ringer solution (500 mL), pentastarch (Pentaspan; 500 mL), and mannitol (Osmitrol; 500 mL) with 5000 IU of heparin to a total volume of 1.5 L. Animals were then randomized to receive either conventional cold-blood cardioplegia (CC) or biologically variable cardioplegia (BVC; water bath temperature = 0°C). All hemodynamic parameters were recorded simultaneously on a 6-channel chart recorder (Gould Instrumentation Systems, Valley View, Ohio) and data-acquisition system (Dataq, Codas). Echocardiography, LVP, and sonomicrometry crystal data were monitored continuously on-line to the Sonometrics computer screen and intermittently recorded for off-line analysis.
After heparinization to an activated clotting time of greater than 400 seconds, CPB was initiated with a flow rate of approximately 3 L/min. We used a Cobe apulsatile roller pump, membrane oxygenator (Cobe Optima Hollow Fiber Open System; Cobe Cardiovascular, Inc, Arvada, Colo), and arterial in-line filter (Baxter 1040D 40-µm pore; Baxter Healthcare Corp, Santa Ana, Calif) with alpha-stat pH management. Normothermic bypass was maintained for approximately 15 minutes to ensure stable hemodynamics and blood gases before placement of the aortic crossclamp. No vasoactive drugs were used during the bypass run. The animals were then actively cooled to 33°C.
Protocol
Control measurements were taken after cannula placement and before initiation of CPB. After aortic crossclamping, cardioplegia was standardized to 20-minute cycles consisting of 5 minutes of either CC or BVC cardioplegia at a mean perfusion pressure of 75 mm Hg and 15 minutes of no perfusion. Five cardioplegia cycles resulted in a total crossclamp time of 85 minutes. We used 1 minute of high-concentration KCl solution (80 mEq/L) and 4 minutes of low-concentration KCl solution (30 mEq/L) at a ratio of 4:1 blood/cardioplegic solution as the perfusing solution.
The computer controller and software for biologically variable control of the roller pump have been described previously. 11 In brief, the systolic variations in arterial blood pressure obtained from a lightly anesthetized, spontaneously breathing pig were used to create the variability file. Minimum, maximum, and mean values for each peak systolic arterial pressure were calculated. The minimum pressure was subtracted from each pressure, and the new minimum (0 mm Hg), maximum, and mean peak values were recalculated and displayed. After initiation of cardioplegia, the target mean cardioplegia pressure and maximum amount of computer modulation were set. We chose a mean perfusion pressure of 75 ± 25 mm Hg pulse variation generated by the computer controller. The roller-pump flow rate was adjusted to achieve the minimum pressure of 50 mm Hg, with the computer modulation level preset to zero. The modulation level was then adjusted until the increase in pump flow rate resulted in an increase in cardioplegia pressure from 50 to 100 mm Hg, the maximum computer modulation level chosen. Computer control of the roller pump was now initiated. At any time, independent setting of the rate control could adjust mean cardioplegia pressure to the predetermined level if required. This setup was complete within 1 minute of initiation of cardioplegia in all cases. Figure 1 shows the pressure profile obtained in the aortic root with either CC or BVC over 35 seconds.
|
Arterial blood gases, hemodynamics, temperatures. and cardiac function were determined at 15, 30, and 60 minutes after CPB. The animals were then killed with an injection of sodium pentobarbital (Euthanol; 100 mg/kg).
Arterial blood gas, temperature, and hemodynamic measurements
At each measurement period, arterial and mixed venous blood gases were sampled, and temperature was recorded. Heart rate, mean arterial pressure, mean pulmonary artery pressure, left ventricular end-diastolic pressure (LVEDP), and left ventricular end-diastolic volume (LVEDV) were measured, and cardiac output was determined.
Cardiac function determinations
At each measurement period, LVEDV was standardized visually to control levels with transfusion. A recording of 15 to 20 heart beats was made at end-expiration. The balloon tip of the inferior vena cava catheter was then inflated for approximately 30 seconds to transiently decrease left ventricular filling so that a range of LVPs at decreasing left ventricular dimensions could be obtained.
Analysis
Postexperiment analysis of the cardiac function curves was done with a Sonometrics software package. For each tracing, left ventricular volume (LVV) was calculated from the crystal pairs by using a 2-dimensional ellipsoid model and plotted against LVP.
Systolic indices
First, the slope of the end-systolic pressure-volume relationship (Emax) was determined from the cardiac function curves obtained over successive beats of decreasing LVV during inferior vena caval occlusion. End-systole was defined as the point in the cardiac cycle at which elastance (LVP/LVV) was maximal. Emax was determined by means of linear regression.
Second, preload recruitable stroke work index was determined over the same cardiac cycles of decreasing LVV. Stroke work was calculated as the integral of each pressure-volume loop and plotted against the corresponding end-diastolic volume, the slope of which represents preload recruitable stroke work index.
Diastolic indices
First, the time constant for ventricular relaxation (
) was calculated from the decline in the LVP pressure trace at stable LVEDP and end-expiration by the following equation:
Pt = P0e(–t/),
where Pt is LVP at any time t, and P0 is LVP at time 0. P0 was determined at minimal dP/dt, the start of isovolumic relaxation. The relationship was determined with a floating asymptote, and was calculated as the negative reciprocal of the slope of the following relationship:
lnPt = –t/ + lnP0
Second, LV diastolic stiffness was quantitated by the stiffness constant (Kp) according to the following equation:
LVEDP = Ce(Kp x LVV)
where C is a derived constant, Kp is chamber stiffness, and LVV is the volume corresponding to a given end-diastolic pressure.
Fractal model of cardioplegia administration
The pressure profile of cardioplegia administration (measured at the cannula site) was analyzed post hoc in both groups. Mean peak systolic arterial pressure (SAP) was obtained in each experiment. Instantaneous pressure measurements (4-ms intervals) were then determined from a representative experiment in both groups. Data were compared in the following manner: instantaneous pressure was subtracted from mean peak SAP, and this value was squared and then log transformed. These data were partitioned into incremental bins of equal size to determine their frequency distribution. The probability of each frequency was determined as Ni/N, where Ni is number of observations in a given frequency bin, and N is total number of observations. A log transformation of the probability distribution was derived. The log probability distribution versus log SAP variation was plotted for the experiment in each group. The equation and correlation coefficient of the data were derived by means of regression analysis.
Cardioplegia pressure-flow model
The relationship between instantaneous pressure and flow was modeled on the canine results of Aldea and colleagues. 5 Regional myocardial pressure-flow relationships fit a family of power functions of the form Q = Px, with x varying between 2.07 to 3.42. We calculated mean cardioplegia flow by averaging instantaneous flow from instantaneous pressures, using these 2 extremes of perfusion. Thirty-five seconds of data from one experiment in each group yielded 8750 instantaneous pressure measurements for this calculation. The BVC/CC mean flow ratio was compared at a pressure normalized to 1.0 for each regional flow. Flow was then calculated at 20%, 40%, and 50% of the normalized pressure to mimic the decrease in perfusion pressure that may occur with increasing downstream closing pressure during cardioplegic arrest.
Comparison of coronary sinus with arterial lactate concentration
We conducted 4 additional experiments in an identical model to compare coronary sinus with arterial lactate concentration before and after CPB. A catheter was inserted through the right atrium deep into the coronary sinus by means of direct palpation and confirmed at autopsy. Cardioplegic solution was administered as in group CC. Simultaneous coronary sinus and arterial blood samples were obtained at baseline and after CPB. Lactate concentrations at the 2 sites were compared with linear regression.
Statistics
Hemodynamic and systolic and diastolic cardiac indices were compared at baseline and at each time period after CPB by a split-plot 1-way analysis of variance for repeated measures. When significant group or group-by-time interactions were found, least squares means test matrices were generated, and within- and between-group comparisons were made. The Bonferroni correction for repeated measures was applied to within-group comparisons. Data are presented as means ± SD unless otherwise noted.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Temperature, arterial blood gases, and serum lactate levels
Nasal and venous temperatures and end-tidal isoflurane concentrations were stable, and arterial blood gases were within the normal range at all measurement periods (Table 1). Temperature decreased equally during CPB (32.9°C for the BVC group and 33.2°C for the CC group), as per the protocol, and was restored to slightly hypothermic levels in both groups after bypass (35.5°C-36°C). Intramyocardial temperature fell to approximately 8.5°C during cardioplegia and rewarmed to approximately 25°C between runs, and this was not different between groups. Similar serum lactate levels were present at baseline and after 60 minutes of CPB. After CPB, serum lactate levels were higher in group CC at every measurement period (group-by-time interaction, P = .008; Figure 2). Comparison of 30 measurements of coronary sinus to arterial lactate concentrations in the 4 post hoc experiments described showed a highly significant correlation (R2 = 0.97).
|
|
|
represents only 35 seconds of the total data set of 38 minutes for BVC. The SD was 11.7 in group BVC and 8.4 in group CC, resulting in an SD ratio of 1.4 in this short sequence. The mean pulse pressure was 26 ± 3 mm Hg in group CC as a result of the roller pump alone. For the entire data set, the imposition of the BVC signal resulted in a 2.5-fold increase of this pulse pressure to 62 ± 10 mm Hg. Volume of cardioplegic solution was not different between groups. Total volumes of high K+ solution delivered were 252 ± 32 mL in group CC versus 248 ± 14 mL in group BVC. Total volumes of low K+ solution were 850 ± 134 mL in group CC versus 922 ± 98 mL in group BVC.
Cardiac events
Five animals experienced ventricular fibrillation on crossclamp release in each group. Electrical defibrillation was required 12 times in group CC and 16 times in group BVC (not significantly different by the Fisher exact test).
Systolic function
Emax was not different at baseline in group CC compared with that in group BVC. The group-by-time interaction was not significant (P = .26), although the absolute value was slightly higher in group CC at all measurement periods after bypass (Table 3). Within groups, there was no significant difference in Emax at any time period compared with baseline values. Preload recruitable stroke work index was equivalent in both groups at baseline and remained unchanged at every measurement period, with no significant group-by-time interactions.
|
). Kp showed marked differences; it was unchanged at any time period in group BVC but doubled in group CC (P = .003) in the immediate postbypass period and remained significantly elevated at every measurement period after bypass between groups (Figure 3). A significant correlation between the change in diastolic stiffness versus the change in serum lactate level after bypass is present in group CC but absent in group BVC (Figure 4).
|
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Diastolic dysfunction is an early indicator of myocardial injury and may occur before changes in contractility. Impaired diastolic dysfunction correlates to difficulty in weaning from CPB and the need for more inotropic support perioperatively. 17 Increases in diastolic stiffness with standard cardioplegia have been well documented previously, 18-20 supporting our findings. All 8 animals in group CC demonstrated a sustained increase in Kp after CPB. With LVEDV controlled between measurement periods, LVEDP increased from 9 to 11 mm Hg after bypass in group CC, which is further corroboration of a noncompliant ventricle. Neither LVEDV nor pressure were different after bypass in group BVC.
Buckberg 21 has integrated multiple cardioprotective strategies, including periods of warm- and cold-blood cardioplegia, antegrade and retrograde delivery, and intermittent and continuous perfusion with both blood and crystalloid solutions to improve myocardial protection. A strategy to maximize distribution of cardioplegic solution on the basis of pump-pressure profiles was not considered. We chose a single approach to cardioplegia administration as a model to determine whether the pressure profile for delivery of cardioplegic solution was an important independent parameter. The results support this contention.
Possible mechanisms to explain the benefit of BVC in this setting include better matching of fractal delivery to a fractal vascular bed, facilitation of vascular recruitment, prevention of microvascular closure, and improved flow generated with a noisy pressure profile.
Fractal model of cardioplegia delivery
Fractal delivery of cardioplegia results in a more variable perfusion pressure, although mean perfusion pressure is identical.Figure 5 shows this difference. When analyzed with the same yardstick of variation (ie, bin sizes of 100 mm Hg2), BVC demonstrated a much noisier signal than that seen with CC. Previous work suggests that collapsed vessels may be better recruited with a noisier input signal. 14 Suki and colleagues 16 have suggested that biologically variable life support is an example of stochastic resonance, enhanced output through use of a noisy input signal. Cardioplegia delivery demonstrates nonlinear pressure-flow relationships. 5 Under such circumstances, a noisier pressure signal results in greater flow because the decrease in flow seen at lower pressures is more than compensated by the increase in flow seen at greater driving pressures. The modeling based on the curves by Aldea and colleagues 5 bears this out with greater calculated flow at the same mean pressure in group BVC. Importantly, a greater advantage is demonstrated as critical closing pressure increases (Figure 6).
BVC and critical closure
Chen and colleagues 22 have demonstrated endothelial swelling, resulting in luminal narrowing in myocardial capillaries after cardioplegic arrest. Such changes increase the propensity for vessel closure. Once closed, large increases in pressure are necessary to reinstitute flow. Failure to recruit vascular territories with each administration of cardioplegia would worsen the ischemic insult, creating a positive-feedback loop for myocardial injury. BVC may help prevent this positive-feedback cycle. From the model, as critical closing pressure increases (perfusion pressure 50% of baseline value), the BVC/CC flow ratio increases(Figure 6). Thus BVC may yield a higher flow at a given perfusion pressure to offer increased myocardial protection. Coggins and colleagues 4 have shown that at 70 mm Hg, 100% of left ventricular areas have significant flow reserve, whereas at 30 mm Hg, equivalent to an increase in critical closing pressure of 40 mm Hg, only 8% of regions had flow reserve. Ghaleh and coworkers 23 showed that areas of high baseline flow, correlating to regions with the highest power-law exponents, are at a 2.3-fold increased risk of myocardial necrosis after 90 minutes of coronary occlusion in the baboon. Thus those regions at greatest risk of ischemia are the same regions that the flow model inFigure 6 would predict are best protected during BVC.
Limitations of the present study
The present study was not designed to measure either microvascular perfusion directly or metabolic markers of coronary perfusion, such as coronary sinus lactate levels. However, evidence of improved coronary perfusion with BVC is suggested indirectly by the diastolic function and arterial lactate results obtained. The increase in serum lactate levels in group CC cannot be explained by a systematic bias in whole-body perfusion. Hemodynamics were well matched between groups at baseline and during CPB, with similar CPB flow rates and serum lactate levels at 60 minutes of CPB. There was no evidence of impaired systemic perfusion in the postbypass period because heart rate, mean arterial pressure, cardiac index, arterial PO2, mixed venous PO2, and urine output were not different between groups. With no evidence for impaired peripheral perfusion and a strong correlation between the change in diastolic stiffness and serum lactates, as seen inFigure 4, the results suggest that cardiac metabolism contributed to the difference in lactate concentrations seen. Follow-up work indicated a high correlation between arterial and coronary sinus lactate levels in this experimental model, suggesting that the above conclusions are valid.
We believe that the present study is the first to demonstrate a benefit of increased pulse variability during cardioplegia delivery. Others have shown improved myocardial function with whole-body pulsatile perfusion after various periods of global myocardial ischemia. 24-28 Although these findings support our conjecture, others have shown no protective effect with whole-body pulsatile perfusion. 27 The present study in pigs is not directly analogous to the clinical situation. The subjects were healthy young pigs, with no evidence of coronary artery disease or left ventricular dysfunction. As such, we anticipated effects to be small. This was borne out by a reproducible effect on diastolic stiffness alone with a crossclamp time of 85 minutes and intermittent cardioplegia, allowing ventricular activity to recur between cardioplegia runs. However, the majority of patients undergoing bypass have myocardia at greater risk for incomplete or inhomogeneous protection during aortic crossclamping and cardioplegic arrest because of critical coronary stenoses, ventricular dysfunction, and hypertrophy. These preliminary results suggest that administering cardioplegia in a fractal manner by using biologically variable pulsation may enhance myocardial protection.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. J. Collins, S. L. Moainie, B. P. Griffith, and R. S. Poston Preserving and evaluating hearts with ex vivo machine perfusion: an avenue to improve early graft performance and expand the donor pool. Eur. J. Cardiothorac. Surg., August 1, 2008; 34(2): 318 - 325. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. D. Mazer, F. Briet, K. R. Blight, D. J. Stewart, M. Robb, Z. Wang, A. M. Harrington, W. Mak, X. Li, and G. M.T. Hare Increased cerebral and renal endothelial nitric oxide synthase gene expression after cardiopulmonary bypass in the rat J. Thorac. Cardiovasc. Surg., January 1, 2007; 133(1): 13 - 20. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. K. Singal, L. M. Docking, L. G. Girling, M. R. Graham, P. W. Nickerson, B. M. McManus, A. B. Magil, E. K.-Y. Walker, R. K. Warrian, M. S. Cheang, et al. Biologically variable bypass reduces enzymuria after deep hypothermic circulatory arrest. Ann. Thorac. Surg., October 1, 2006; 82(4): 1480 - 1488. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. Goldberger Giles F. Filley Lecture. Complex Systems Proceedings of the ATS, August 1, 2006; 3(6): 467 - 471. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. F Brewster, M. R. Graham, and W. A. C Mutch Convexity, Jensen's inequality and benefits of noisy mechanical ventilation J R Soc Interface, September 22, 2005; 2(4): 393 - 396. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. BOKER, M. RUTH GRAHAM, K. R. WALLEY, B. M. MCMANUS, L. G. GIRLING, E. WALKER, G. R. LEFEVRE, and W.A. C. MUTCH Improved Arterial Oxygenation with Biologically Variable or Fractal Ventilation Using Low Tidal Volumes in a Porcine Model of Acute Respiratory Distress Syndrome Am. J. Respir. Crit. Care Med., February 15, 2002; 165(4): 456 - 462. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
