| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;80:686-694
© 2005 The Society of Thoracic Surgeons
a Division of Cardiothoracic Surgery, University of Miami Miller School of Medicine, Miami, Florida
b Division of Pediatric Cardiology, University of Miami Miller School of Medicine, Miami, Florida
c Department of Neurological Surgery, University of Miami Miller School of Medicine, Miami, Florida
Accepted for publication March 4, 2005.
* Address reprint requests to Dr Ricci, Division of Cardiothoracic Surgery, Daughtry Family Department of Surgery, University of Miami, Holtz Center 3072 (R-114), 1611 NW 12th Ave, Miami, FL33136 (Email: mricci{at}med.miami.edu).
Presented at the Forty-first Annual Meeting of The Society of Thoracic Surgeons, Tampa, FL, Jan 24–26, 2005.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
MATERIAL AND METHODS: Univentricular physiology was created in 8 piglets, while 8 animals were sham controls. Aortopulmonary shunt, echocardiography-guided atrial septostomy, tricuspid valve avulsion, and pulmonary artery occlusion were performed to allow the left ventricle to support systemic and pulmonary circulations. Cerebral blood flow (microspheres), cerebral O2 and lactate metabolism, and cerebral O2 saturation were measured at baseline, 30 minutes, and 120 minutes after conversion to univentricular physiology.
RESULTS: Cerebral blood flow increased in the cerebrum and subtentorium in controls (p < 0.05), whereas it remained unchanged in univentricular piglets. Cerebral O2 delivery at 30 and 120 minutes was lower in univentricular physiology than in controls (p = 0.05). Fractional oxygen extraction was unchanged in both groups. Cerebral O2 consumption trended lower in univentricular physiology (p = not significant), while it was unchanged in controls. Lactate cerebral metabolic rate (CMRLactate) increased at 30 and 120 minutes in both groups. The decline in O2 delivery was variable, but present in nearly all brain regions.
CONCLUSIONS: This study confirms the hypothesis that, in univentricular physiology, hypoxemia and limited cerebral blood flow reduce cerebral O2 availability in nearly all regions. These findings contribute to the understanding of brain abnormalities in infants with univentricular physiology.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
The etiology of neurologic deficits in infants with congenital heart disease is incompletely understood, and is likely to be multifactorial [3]. Recent clinical evidence points to the combination of hypoxemia and low cardiac output in the early neonatal period as a risk factor for periventricular leukomalacia [2]. This encompasses a spectrum of lesions characterized by multiple areas of necrosis in the periventricular white matter [4]. The pathophysiology of this entity, often found in premature infants, has been shown to involve hypoxia/ischemia [4]. Further, it has been shown that, in infants with complex congenital heart defects, preoperative cerebral blood flow (CBF) is often diminished, and that lower levels of CBF are associated with periventricular leukomalacia [5].
These clinical observations suggest that the combination of reduced CBF and hypoxemia is strongly associated with cerebral abnormalities such as periventricular leukomalacia, microcephaly, and cortical atrophy [6]. Infants with SVP could be especially vulnerable to brain injury as they are exposed to hypoxemia and have a limited cardiac output reserve. Despite the growing clinical evidence, these assumptions have never been tested in the animal laboratory. Only a few animal models of SVP have been reported in the literature [7–12]. These models have focused mainly on strategies to optimize blood flow distribution between systemic and pulmonary circulations, and on the role of pharmacologic and respiratory interventions in altering this balance [8–12]. To date, experimental data regarding the consequences of SVP on cerebral blood flow and regional cerebral O2 availability are lacking. The purpose of this study was to test the hypotheses that in acute SVP (1) the increase in CBF induced by the lower arterial O2 content is inadequate to maintain cerebral O2 delivery, (2) the ability of the brain to increase O2 extraction is impaired owing to hypoxemia, and (3) cerebral O2 delivery may be diminished selectively in different areas of the brain.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Surgical Preparation
Piglets were anesthetized with intramuscular ketamine (40 mg/kg) and xylaxine (4 mg/kg), intubated by tracheostomy, and placed on volume-control ventilation (tidal volume 25 to 30 mL/kg, rate of 25 breaths per minute, inspired O2 fraction of 0.25). Anesthesia was maintained with fentanyl (50 mcg·kg–1
·h–1), pancuronium (0.4 mcg·kg–1
·h–1), and midazolam (0.2 mg·kg–1
·h–1). A catheter was inserted in the femoral artery for pressure monitoring and blood sampling. A 6F introducer sheath was placed in the femoral vein for fluid administration and subsequent insertion of a balloon atrial septostomy catheter. A catheter was inserted in the left internal jugular vein for venous blood sampling. In some animals (n = 9), a catheter was also inserted in the sagittal sinus through a midline burr hole in the skull. Electrocardiography and temperature monitoring were obtained.
The single ventricle model resembled that described by Randsbaek and colleagues [7]. Through a median sternotomy, catheters were placed in the right and left atrium. Heparin was given (150 units/kg), and a 3.5-mm Gore-Tex shunt was interposed between the aorta (take-off of the innominate artery) and the pulmonary artery. While the shunt was clamped, a 2 mL balloon septostomy catheter (B. Brawn Medical, Bethlehem, PA) was advanced from the right femoral vein into the right atrium. Epicardial two-dimensional echocardiography was used to direct the catheter across the atrial septum and perform a pull-back septostomy. The same catheter was then advanced into the right ventricle. The tricuspid valve was made incompetent by repeatedly withdrawing the inflated balloon across the valve. Lastly, the shunt was opened, and the main pulmonary artery was occluded. This allowed the left ventricle (LV) to support both systemic and pulmonary circulations. Of importance, this is a difficult model, and a failure rate of approximately 60% to 70% can be expected. Most failures are due to the absence of a patent foramen in the atrial septum, which makes the atrial septostomy impossible.
Animals treated as sham-controls received identical surgical preparation. After the sternotomy, the shunt was constructed and clamped. The atrial septostomy and tricuspid valve avulsion were simulated by manipulating the catheter against or across the atrial septum and tricuspid valve, without inflating the balloon. Piglets were kept in a biventricular state.
Experimental Protocol
Eight piglets were included in the study group (SVP), and 8 were treated as sham-operated controls. In both groups, animals were ventilated at a constant FiO2 of 25%, adjusting the respiratory rate to maintain a partial pressure of carbon dioxide (pCO2) between 35 and 45 mm Hg. Rectal temperature was kept at 35.5° to 36.5°C. Based on experience from pilot experiments, a standard regimen was developed to facilitate establishing our SVP model. Throughout the experiments, normal saline (4 cc·kg–1
·min–1), dopamine (5 to 10 mcg·kg–1
·min–1), and epinephrine (0.05 to 0.1 mcg·kg–1
·min–1) were administered in all animals to maintain cardiovascular stability and to avoid hemodynamic perturbations. Dopamine administration did not differ between the SVP and control group (from 7.1 ± 0.8 to 7.8 ± 0.8 to 6.5 ± 0.8 µg·kg–1
·min–1 in SVP animals at baseline, 30 minutes, and 120 minutes, respectively; versus 8.5 ± 0.7 to 8.5 ± 0.7 to 8.5 ± 0.7 µg·kg–1
·min–1 in controls; p = 0.7), nor did epinephrine administration (from 0.07 ± 0.01 to 0.07 ± 0.01 to 0.06 ± 0.01 µg·kg–1
·min–1 in SVP animals at baseline, 30 minutes, and 120 minutes, respectively; versus 0.065 ± 0.008 to 0.06 ± 0.008 to 0.06 ± 0.008 µg·kg–1
·min–1 in controls; p = 0.9). Calcium gluconate and sodium bicarbonate were given as needed. Fresh whole blood was obtained from an adult pig and infused to maintain a hemoglobin concentration close to baseline. The same protocol was used in the control group.
Physiologic measurements were obtained at baseline and at 30 and 120 minutes after conversion to SVP. These included hemodynamic parameters, blood sampling (arterial, central venous, and cerebral venous, hemoglobin, and lactate), determination of total cardiac output by electromagnetic flowmeter, and determination of regional cerebral blood flow by stable-isotope microspheres injections. At completion, piglets were sacrificed with KCl and fentanyl. An autopsy was performed to confirm the correct positioning of all indwelling catheters and the adequacy of the atrial septostomy. The brain was weighed and dissected to obtain samples from the cortex, white matter, thalamus, hippocampus, cerebellum (vermis), medulla, pons, and mesencephalon for regional blood flow determinations.
In order to establish the effect of the surgical maneuvers on the parameters studied, we performed sham-operated controls (n = 8). These animals received identical surgical preparation. After baseline measurements, subsequent measurements were obtained after the simulated septostomy, with the animals in a normal biventricular state, at 30 minutes and 120 minutes.
Regional Cerebral Blood Flow Determinations
Stable-isotope microspheres (15 ± 5 µm [BioPhysics Assay Laboratory, Worcester, Massachusetts]) were used as described by others [13]. Microspheres labeled with nonradioactive isotopes were administered over 3 seconds through the left atrial catheter, while reference samples were collected over 90 seconds from the femoral artery (2 mL/min). A total of 1 x 106 microspheres was injected in the biventricular state, whereas 2.5 x 106 were injected in SVP. Experience with microspheres in a univentricular model is lacking in the literature. Based on previous pilot experiments, we demonstrated that 2.5 x 106 microspheres in the univentricular state compensated for run-off through the pulmonary circulation and ensured adequate microsphere concentration in systemic organs. At the end of the experiment, tissue samples were weighed fresh, and dried in a warming-oven at 60°C for 48 hours. Dried samples were sent for analysis and processed by "neutron activation" [13]. This technique entails exposure of the isotope to a neutron beam, which results in an activated radioactive nucleus. Activated labels are then allowed to decay for 48 h, during which time gamma-rays are emitted and measured. The signal is proportional to the total mass of the isotope, and therefore to the concentration in the sample [13]. Results of the assay for each label are reported in disintegrations per minute (dpm). Cerebral blood flow is expressed in mL·min–1
·100 g–1, and calculated by normalizing the concentration of microsphere in the tissue sample (dpm/g) to the concentration in the reference sample (dpm x minutes x mL–1).
Regional Cerebral Oxygen Saturation
A commercially available device (INVOS 5100 Cerebral Oxymeter; Somanetics, Troy, Michigan) was used. A 30-mm patch containing the near infrared spectroscopy (NIRS) emitting source and sensor was placed on the forehead of each animal and connected to the infrared spectrophotometer unit.
Physiologic Measurements and Calculations
Total cardiac output (excluding coronary blood flow) was determined by using an electromagnetic flowmeter (Transonic Systems, Ithaca, New York) placed on the ascending aorta. After conversion to SVP, pulmonary blood flow was determined by placing the flowmeter on the pulmonary artery distal to the aortopulmonary shunt [7]. Total systemic blood flow was calculated as follows: total cardiac output (mL/min) - pulmonary blood flow (mL/min).
Oxygen delivery (mLO2/mL) was calculated as blood flow (mL·min–1 ·100 g–1) x arterial O2 content. Arterial O2 content (CaO2) was calculated as (Hb x 1.36 x SaO2) + (0.0031 x PaO2). Oxygen consumption was calculated as arterial minus venous oxygen content (blood obtained from sampling the jugular vein). Oxygen extraction was calculated as oxygen uptake divided by oxygen delivery. To establish the validity of O2 calculations based on measurements from the jugular vein, we compared calculations obtained by using jugular venous blood versus sagittal sinus blood in a limited number of animals (n = 9).
Statistical Analysis
All data were expressed as means ± SEM. Comparisons within each group and between groups were made using one-way and two-way analysis of variance (ANOVA). Tukey post hoc testing was used to detect statistically significant differences among the study periods and be-tween groups at specific data points. A commercially available software package was used (SigmaStat2.0). A p value of less than 0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
|
|
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
In contrast to Stonestreet and colleagues [16], a major new finding of this study is that when hypoxemia is introduced acutely as part of SVP, the increase in CBF is attenuated, and is inadequate to maintain cerebral O2 delivery. In fact, in our SVP model, cerebral O2 delivery declined by 25% at 30 minutes and remained low at 120 minutes, whereas it increased by 70% in controls (p = 0.05) as a result of the increase in CBF probably caused by the surgical stress and inotropic support. Owing to the lower arterial O2 content in SVP, it was logical to expect to see the highest increase in CBF in SVP rather than controls. Also concerning was that in SVP the brain failed to increase O2 extraction despite the declining O2 delivery and consumption. While the reasons for this are unclear, our observations are consistent with those of others [17], suggesting that the mechanisms by which the brain increases O2 extraction may be impaired during hypoxemia. Cerebral O2 delivery is determined by the interaction between CBF and arterial O2 content. Under normal conditions, the brain receives fully saturated blood and can increase O2 extraction to sustain metabolic demands, if CBF declines. Conversely, our findings suggest that in the presence of hypoxemia the brain looses its ability to adequately increase O2 extraction. Therefore, in SVP cerebral O2 delivery is highly dependent on CBF, and a decline in CBF may directly compromise cerebral O2 availability.
In our model we did not test the effect of hemodilution on cerebral O2 delivery. In the SVP group, CBF increased inadequately and cerebral O2 delivery declined despite the hemoglobin was unchanged. Previous studies have shown that neurologic recovery is enhanced by a higher hematocrit following hypothermic circulatory arrest [18]. Todd and coworkers [17] have shown that CBF correlates inversely with CaO2. However, the increase in CBF is much greater when low CaO2 is caused by hypoxemia rather than hemodilution [17]. Therefore, it could be speculated that hemodilution in the setting of hypoxemia would not further increase CBF as compared with hypoxemia alone, and instead might further compromise O2 delivery. A corollary is that higher levels of hemoglobin could be critical to maximize cerebral O2 delivery in an environment of chronic hypoxemia and poor CBF reserve, such as SVP.
Our findings showed that cerebral O2 consumption declined by 27% at 120 minutes in SVP, whereas it remained unchanged in controls. Previous studies have shown that cerebral ischemic injury is unlikely unless cerebral O2 consumption declines by more than 40% to 50% [19, 20]. The neonatal brain has a series of conservation strategies such that smaller reductions are unlikely to produce neuronal damage [21]. Therefore, our data does not suggest that cerebral ischemia had occurred in SVP, as also indicated by the raise in cerebral lactate utilization, rather than production. However, it is concerning that such a trend was observed concomitantly with declining O2 delivery and O2 fractional extraction. The increase in lactate consumption noted in both SVP and controls is difficult to interpret, as it could have resulted from stress, decreased O2 availability, or higher circulating lactate [22].
Regarding CBF heterogeneity, the increase in CBF after conversion to SVP was greater in the cerebrum than in subtentorial regions at 30 minutes (Fig 1). However, in both the cortex and the white matter, CBF returned to baseline, or below baseline, at 120 minutes. Only in the hippocampus was the increase in CBF sustained at 120 minutes (Fig 1), resulting in unchanged O2 delivery (Fig 2). This is consistent with previous experiments showing the ability of this area to increase blood flow profoundly in response to hypoxia [15]. Regarding O2 delivery, increases on the order of 30% to 70% were seen in all regions in controls, while a 10% to 30% decline was observed in SVP, except in the hippocampus. The decline was most severe in the cortex and white matter, where O2 delivery continued to decrease at 120 minutes (Fig 2), although it was also notable in other regions such as the cerebellum and the subtentorium as a whole. It is possible that in our experiment the decline in O2 delivery to cortex and white matter was underestimated. It has been shown that watershed areas are at risk for hypoxia/ischemia as they have lower CBF [23]. Owing to the small size of the piglet brain and to the need for providing sufficient tissue for microsphere analysis, samples from the cortex and white matter were not taken selectively from watershed areas but were obtained from the entire hemisphere.
Although significant emphasis has been recently placed on the white matter and cortex, our findings support previous observations that the pattern of injury may be diffuse [2, 5, 6]. Despite SVP may cause more severe oxygen deprivation in certain areas of the forebrain, our findings suggest a global state of reduced cerebral O2 availability that, although subtle, may be sustained. In clinical SVP, overt cerebral hypoxia-ischemia could be precipitated by acute situations of reduced cardiac output and CBF. Also, factors that cause a reduction in arterial O2 content (hemodilution, severe hypoxemia) are likely to lower further the threshold for brain hypoxic injury due to the limited CBF reserve. It has also been shown that different regions of the brain may have different susceptibility to hypoxic insults at various maturational stages [24]. Therefore, selective injury to certain areas of the brain could also depend on the timing of the insult [25]. The recurrent pattern of injury involving the white matter could be due to the selective susceptibility of immature oligodendrocytes to hypoxia [26].
In our experiment, we utilized piglets as this species has been used previously for SVP models, and the patterns of CBF and O2 metabolism are well characterized. We chose to perform the model without bypass to eliminate the effects of an additional variable on the parameters measured. However, this study presents several limitations, some of which are characteristic of translational physiology experiments. It is unknown whether these findings can be extrapolated to the clinical setting. Also, as SVP was created acutely in animals with normal cardiovascular physiology, it may not accurately reproduce clinical SVP. Additional limitations relate to the relatively short period of observation due to rapid model deterioration.
Despite these limitations, this study suggests that in acute SVP the combination of hypoxemia and limited CBF reduces cerebral O2 availability. Further, it provides a physiologic basis to the understanding of brain abnormalities in infants with SVP.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
My question relates to the decrease in the cerebral blood flow that you see with the model. Do you think that it could be related to the shunt? Is there a steal? Just like our concern with myocardial blood flow in the typical Norwood with the BT shunt; do you think that the shunt has anything to do with the decreased cerebral blood flow? And would you be willing to take that model to another level and see if you could bring the pulmonary blood flow from the left ventricle rather than from the aorta with the shunt?
DR RICCI: This model, as you said, is very difficult and also very humbling. Our first thought was that there could be a difference between, for example, right and left carotid flow related to the shunt. We have measured in all animals, although the data was not presented here, at each of the data point, carotid blood flow by using a transonic flow meter. We have not found any difference.
However, it needs to be said that the arch anatomy of pigs is different from humans. Pigs have some sort of common trunk that originates from the aorta, and the bifurcation between right and left common carotid arteries is very proximal.
Our model was slightly different from yours because we did not divide the innominate artery. We just partially clamped the aorta and the innominate at the takeoff and then we based the shunt there. So the shunt could have, in order to answer your question, affected blood flow both to the right and left, and maybe a difference was not notable.
We also used a different flow meter to look at diastolic and systolic flow curves, with the assumption that if the shunt was responsible for flow disturbance, one could see a variation in flow curve with possibly diastolic reversal in the carotids. However, we did not see that either, although we have these data only in a few animals.
In order to answer your second question, we’re now working on a model of ventricle-to-pulmonary artery shunt, which is also a very challenging model. This model, as you know, was done without cardiopulmonary bypass, because we thought that by avoiding this element of variability one could have a better understanding of the changes in cerebral blood flow and oxygen delivery that are truly due to the change in physiology rather than to other variables. So the challenge will be to try to do the ventricle-to-pulmonary artery single ventricle model without using bypass so as to be able to compare data.
DR JOHN J. NIGRO (Los Angeles, CA): I wondered at any point did you try to establish or did you measure the relationship between arterial pCO2 and cerebral blood flow in this model?
DR RICCI: We did not study the effect of variations in CO2. We made every attempt at maintaining CO2 as close as possible to baseline, as our focus was restricted to the change in physiology. Pigs have CO2 like humans in the low 40s, high 30s.
What we did note, was that in single ventricle piglets the CO2 trended higher after single ventricle physiology was established. So it’s kind of intriguing that those declines in cerebral blood flow and oxygen delivery were seen actually in the face of a trend for higher CO2 in single ventricle piglets, which by itself should have increased blood flow to the brain. We did not, however, make any comparisons by changing the CO2. We hope to be able to do that in future experiments.
DR NIGRO: It’s just an interesting issue, maybe permissive hypercapnia would provide better overall cerebral oxygenation as has been demonstrated in the Glenn circulation. It would be interesting to try to document it in this model.
DR J. WILLIAM GAYNOR (Philadelphia, PA): I may have missed it because I couldn’t see all the slides. Did you measure cerebrovascular resistance?
DR RICCI: We did not.
DR GAYNOR: Because it looks like the cerebral circulation vasodilates and regional cerebral blood flow goes up, so there is an attempt to increase cerebral blood flow.
The hemoglobins are much lower than we would normally maintain in a child with single ventricle and hypoxemia. It looks like they were around 9 or 10 for a hemoglobin, which normally we would have a hypoplast with a crit of 45.
Do you think that some of this just represents that there’s a limit that the cerebrovascular resistance can fall and cerebral blood flow can increase, particularly in a shunt, and could you increase your cerebral oxygen delivery to satisfactory levels just by increasing the hemoglobin?
DR RICCI: I cannot answer your question regarding the cerebrovascular resistance, as we did not place an intracranial catheter to measure ICP.
I have to say that hemoglobin was a very intriguing factor. First, what we noted is that piglets have a much lower hemoglobin even at baseline than humans. And that occurred even more so in the control group, in which the starting hemoglobin was lower. For some reason, we received a few animals that had a lower hemoglobin and made our experiment more difficult, because at that point we were unsure as to whether we should have transfused the animals even before doing the baseline measurements to increase the hematocrit, or we should have kept the hemoglobin as close to the physiologic starting point normal as possible. Additionally, the animals lost some blood during the experiment. That cannot be avoided, as it was due to the sternotomy, the surgical manipulations, and the shunt. These maneuvers were carried out in controls as well, so there was a similar blood loss in these animals. We could not have done the experiments without blood.
Having said that, we decided not to use higher levels of hemoglobin by transfusing all animals at the beginning of the experiment because we were not sure as to how that could affect the experiment. We only transfused the animals to keep the hematocrit close to the starting point. However, our speculation is that higher hematocrits could be critical to maintain oxygen delivery in single ventricle physiology. This would need further investigation.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Ricci, P. Lombardi, A. Galindo, S. Schultz, A. Vasquez, and E. Rosenkranz Effects of single-ventricle physiology with aortopulmonary shunt on regional myocardial blood flow in a piglet model. J. Thorac. Cardiovasc. Surg., August 1, 2006; 132(2): 252 - 259.e2. [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 |
