Effects of Aortopulmonary Collaterals on Cerebral Cooling and Cerebral Metabolic Recovery After Circulatory Arrest
Background Aortopulmonary collaterals (APC) have been associated with an increased risk of choreoathetosis after deep hypothermic circulatory arrest (DHCA). To study the effects of APC on cerebral hemodynamics and metabolism before and after DHCA, a piglet model was developed.
Methods and Results Protocol 1: Eight 4- to 6-week-old piglets underwent placement of a left subclavian–to–main pulmonary artery shunt. Control shunts (n=4) were ligated, APC shunts (n=4) were left patent. All animals were placed on cardiopulmonary bypass (CPB) and cooled in identical fashion for 20 minutes. Temperature probes were placed in the nasopharynx, cortex, and deep brain. Control animals achieved significantly lower temperatures in all three areas by the end of cooling (17.5°C versus 20.1°C, 19.0°C versus 22.3°C, and 17.5°C versus 21.0°C, respectively, P<.005). Protocol 2: Six control and six APC animals were instrumented as described. All were placed on CPB, cooled to 18°C, arrested for 90 minutes, and rewarmed to 37°C. Cerebral blood flow (CBF) was measured with radioactive microspheres while warm on CPB, after cooling, and after rewarming. Arterial and sagittal sinus blood gases and CBF were used to calculate the cerebral metabolic rate of oxygen consumption (CMRO2). Both CBF and CMRO2 were significantly higher after rewarming to 37°C in control versus APC animals (28±3 versus 14±2 mL/100 g per minute and 1.72±0.21 versus 1.04±0.14 mL O2/100 g per minute, respectively, P<.05).
Conclusions APC decrease the rate of cerebral cooling on CPB and even if temperature is controlled result in increased cerebral metabolic derangement after DHCA. Patients with such collaterals may need additional measures to optimize cerebral protection.
Hypothermic CPB at 16° to 18°C, with or without DHCA, is commonly used for the repair of congenital cardiac defects. The use of hypothermic circulatory arrest is based on the well-described decrease in cerebral and total body metabolic requirements at such low temperatures.1 2 3 4 Despite this metabolic protection, postoperative neurological events that range from single seizures to catastrophic neurological deficits have been reported in as many as 25% of cases.5 The majority of these events result in no long-term morbidity; however, postoperative choreoathetosis has been associated with significant morbidity and mortality.3 6 7 This movement disorder has been reported after either hypothermic CPB alone7 or with DHCA,3 6 with an incidence that ranged from 1% to 12%.8 Although the etiology of choreoathetosis is unknown, recent studies have demonstrated abnormalities of the cortical and subcortical regions9 and an association with later repair, cyanotic heart disease, shorter cooling periods, and the presence of APC.3
Prior studies have clearly demonstrated the effects of APC on myocardial perfusion10 and the distribution of cardiac output during surface cooling.11 However, there is a paucity of information regarding the effects of such shunts on cerebral perfusion and metabolism during CPB and after DHCA.
The goal of this study was to develop a reliable animal model of pediatric APC for the study of cerebral responses to CPB and DHCA. Such a model would exclude many of the confounding variables and heterogeneity inherent to clinical reviews and thus allow a focused evaluation of the effects of APC.
Twenty 4- to 6-week-old DeKalb piglets weighing 9 to 13 kg were premedicated with ketamine (20 mg/kg IM) and acepromazine (1 mg/kg IM), intubated, and placed on mechanical ventilation (Sechrist Infant Ventilator, model IV-100B). The piglets were anesthetized and paralyzed with fentanyl (100 μg/kg IV bolus and 50 μg/kg per hour infusion) and pancuronium (0.3 mg/kg IV). The ventilator was set with a positive inspiratory pressure of 25 mm Hg and positive end-expiratory pressure of 3 mm Hg. Respiratory rate and inspired oxygen fraction were titrated to maintain an arterial Pco2 of 35 to 45 mm Hg and Po2 of 200 to 300 mm Hg. Sodium bicarbonate (8.5%) was used to maintain a base excess between −3 and 3 mmol/L. All animals received methylprednisolone (25 mg/kg IV) preoperatively.
Shunt Placement and Instrumentation
A femoral arterial cannula was placed for blood pressure and arterial blood gas monitoring. A generous left thoracotomy was then performed through the fourth intercostal space. The upper lobe of the left lung was retracted inferiorly, and the left subclavian artery was exposed. Heparin (500 U/kg IV) was administered, and a 4-mm polytetrafluoroethylene graft was placed between the subclavian artery and the main pulmonary artery (Fig 1⇓). The animals were randomized to either the control group, which had the shunts immediately ligated, or the APC group, which were allowed to have patent shunts.
The animals were then placed in supine position and a median sternotomy was performed. NP (YSI-400, Yellow Springs Instruments), cortical, and deep brain temperature probes (Shiley Inc, 15-mm and 30-mm Myocardial Probes) were placed, and an ultrasonic flow probe was placed on the pulmonary artery proximal to the shunt. Cannulas (10F aortic, 28F right atrial, and 22F left atrial) were placed for CPB and venting of the left atrium. The CPB circuit consisted of two Stockert Shiley roller pumps (model 10-10-00, Shiley Inc), Cobe membrane oxygenators (Cobe Laboratory), and a Biocal 370 heat exchanger (Biomedicus). The pump was primed with whole donor blood.
The first eight animals were randomized into control or APC groups as described. They were placed on CPB at 85 mL/kg per minute and core-cooled for 20 minutes with a heat exchanger temperature of 7° to 9°C. The left atrium was vented into the circuit to prevent pulmonary edema during cooling and rewarming. NP, cortical, and deep brain temperatures were recorded every minute during cooling. Radioactive microspheres were injected warm on CPB and after 5, 13, and 20 minutes of cooling. Microspheres (3×106 in 1 mL 10% dextran and 0.01% Tween) were used for each injection (153Gd, 113Sn, 103Ru, 95Nb, and 46Sc in random order; NEN Research Products, DuPont). During each injection an arterial reference sample was withdrawn at 7 mL/min over 120 seconds starting 10 seconds before the injection (Harvard Apparatus, model 4400-001). Arterial and sagittal sinus blood gases were also drawn.
At the end of the cooling period the animals were euthanatized, and the brain and lungs were harvested. The brain was subdivided into the cortex, basal ganglia, cerebellum, and brain stem. All tissue and arterial reference samples were counted in a gamma counter (Minaxi Auto-gamma Counter, 5000 Series, Packard Instrument Co), and individual nuclide activity was calculated (Compusphere Microsphere Multinuclide Analysis Software, Packard Instrument Co).
CMRO2 was calculated as follows:
Blood flow is expressed as mL/100 g per minute and CMRO2 as mL of O2/100 g per minute.
Twelve piglets were randomized into control and APC groups as described. All were placed on CPB as in protocol 1, cooled to 18°C over 20 minutes, arrested for 60 minutes, and rewarmed to 37°C over 30 minutes. APC animals required heat exchanger temperatures of 3° to 7°C to reach an NP temperature of 18°C in 20 minutes. Radioactive microspheres were injected warm on CPB, after cooling, and at the end of the rewarming period. Tissue and blood samples were collected and analyzed as in protocol 1.
Paired Student’s t tests with the Bonferroni correction for repeated measures were used to compare time points within groups and unpaired t tests for comparisons between the two groups. A value of P<.05 was considered significant for intergroup comparisons.
All experiments were conducted with the approval of the institution’s Animal Care and Use Committee, and animals received humane care in compliance with the “Guide for the Care and Use of Laboratory Animals” published by the National Institutes of Health (NIH Publication No. 85-23, revised 1985).
Cerebral Temperature and Blood Flow During Cooling—Protocol 1
Temperatures in NP, cortex, and basal brain were compared at baseline and at the end of cooling. There was no significant difference between groups at baseline, but the control animals cooled more quickly, resulting in significantly higher temperatures in the APC group by the end of cooling (Fig 2⇓).
Pulmonary and global cerebral blood flows expressed in mL/100 g per minute and cerebral oxygen extraction are summarized in Table 1⇓. While on CPB, the right ventricle was visually checked for proper emptying, and zero flow through the proximal pulmonary artery was confirmed with the ultrasonic flowmeter; therefore, all pulmonary flow while on CPB was either shunt or bronchial flow. The average pulmonary-to-systemic flow ratio (Qp/Qs) was 1.8:1 prior to CPB, but as the animals were cooled on CPB shunt flow markedly increased from 83±21 to 270±35 mL/100 g per minute with no change in CPB pump flow. This resulted in an increased pulmonary-to-systemic flow ratio in the APC group as cooling progressed. Global cerebral blood flow was significantly decreased in the APC group during cooling. Despite this reduction in cerebral blood flow on CPB and during cooling, the CMRO2 was unchanged. Maintenance of CMRO2 was achieved through a significantly increased cerebral oxygen extraction in the APC group (30.0±2.1% versus 20.2±1.3%, P=.008 at the end of cooling).
Cerebral Blood Flow and Metabolic Recovery After Circulatory Arrest—Protocol 2
Regional and global cerebral blood flows before and after DHCA are summarized in Table 2⇓. Although blood flow to the basal ganglia and brain stem was significantly decreased in the APC animals after cooling, these regions recovered blood flow as well as controls after DHCA. Flow to the cortex and cerebellum on the other hand was markedly decreased after arrest. Since the cortex is considerably larger than the other regions of the brain, the global cerebral blood flow follows a similar pattern with significantly worse recovery in the APC group.
CMRO2 before and after DHCA is displayed in Fig 3⇓. There was no difference in the CMRO2 between the two groups at baseline, and both groups were equally well suppressed by the end of the cooling period. DHCA resulted in a significant decrease in the CMRO2 in both the control and APC animals (3.43±0.40 to 1.72±0.21 mL O2/100 g per minute, P=.026 versus 3.89±0.35 to 1.04±0.14 mL O2/100 g per minute, P=.001). However, the APC group had a significantly lower CMRO2 than control animals after DHCA (1.04±0.14 versus 1.72±0.21 mL O2/100 g per minute, P=.02). The average Qp/Qs in the APC animals was 1.7:1 before CPB.
The presence of large APC has been associated with an increased risk of postoperative choreoathetosis after CPB and circulatory arrest.3 It has been suggested that a cerebral “steal” phenomenon may be responsible for this association. If a cerebral steal is taking place during CPB, it could increase neurological injury in three ways. First, cerebral ischemic injury could occur during warm CPB, while the child is dependent on a fixed pump flow, and cerebral oxygen delivery to the brain may be impaired. During normothermic CPB, the brain has been shown to autoregulate blood flow over a wide range of perfusion pressures to satisfy metabolic needs.14 15 Given this ability, one would expect any episode of hypoperfusion that was severe enough to produce cerebral ischemic injury to be associated with signs of systemic hypoperfusion such as acidosis and multisystem organ injury. Since these signs of systemic insult are rarely present, this mechanism of direct ischemic injury is unlikely.
Second, in the cases of hypothermic and low-flow CPB and DHCA, in which hypothermia is used to protect the brain from intentional decreases in cerebral O2 delivery, relatively smaller decreases in cerebral blood flow could result in heterogeneous cerebral cooling. This could theoretically expose certain regions of the brain to ischemic injury during low-flow CPB or DHCA due to inadequate regional cooling and protection.
Finally, neurological injury could occur during the rewarming period if cerebral blood flow and metabolism are not properly matched. Croughwell et al16 have suggested that the rewarming period could be a particularly vulnerable time for the brain since they found a correlation between jugular bulb desaturation during rewarming and lower postoperative neuropsychological test scores. Van der Linden et al17 also observed significant jugular venous desaturation during rewarming in children after DHCA. Such mismatching of cerebral blood flow and metabolic needs could be exacerbated by aortopulmonary runoff.
The model of APC described here proved reliable and consistent with regard to shunt flow and hemodynamics. Prior studies11 have described the use of a 6-mm graft between the ascending aorta and the main pulmonary artery. While the placement of a shunt in this position is technically simpler, it precludes the use of clinically standard ascending aortic cannulation for CPB. The left subclavian artery was easily accessed through a left thoracotomy and its use kept the ascending aorta available for cannulation. Placement of a 4-mm shunt was well tolerated with systemic-to–pulmonary flow ratios that ranged from 1.5 to 2.2:1. This provided a clinically relevant pediatric shunt model that paralleled the clinical setting as closely as possible.
The findings of this study support the contention that APC can result in inadequate cerebral cooling if additional precautions are not taken. However, even if the brains of animals with APC are cooled to the same temperature as control animals, increased cerebral injury is observed after DHCA. The cause of this injury cannot be clearly defined by these results, but the post-DHCA regional blood flows suggest that the cortex and cerebellum suffered the bulk of the injury. The brain stem and basal ganglia were better able to maintain their blood flows. Unfortunately, regional temperatures and metabolism could not be assessed with these methods.
Given these findings, it would be prudent to maximize cerebral protection in children with APC that cannot be embolized preoperatively or ligated intraoperatively prior to cooling and DHCA. Gillinov et al18 have demonstrated superior neurological protection for normal dogs with profound hypothermia with tympanic membrane temperatures of 5° to 7°C, and Mault et al19 have demonstrated improved cerebral protection with topical cooling and intermittent perfusion. The use of pH stat blood gas management during cooling has been shown to improve cooling efficiency by vasodilating the cerebral vasculature. Further study of these techniques or others, such as increasing CPB pump flow during rewarming, could help to define the optimal protocol for the treatment of these difficult patients.
Selected Abbreviations and Acronyms
|CMRO2||=||cerebral metabolic rate of O2 consumption|
|DHCA||=||deep hypothermic circulatory arrest|
We would like to thank Ronnie Johnson for his expert assistance and dedication to surgical research.
Presented in part at the 80th Annual Clinical Congress of the American College of Surgeons, Surgical Forum, Chicago, Ill, October 9-14, 1994, and The 67th Scientific Sessions of the American Heart Association, Dallas, Tex, November 14-17, 1994, and published in abstract form (Surg Forum. 1994;45:259-260 and Circulation. 1994;90(suppl I):I-253.).
- Copyright © 1995 by American Heart Association
Wong PC, Barlow CF, Hickey PR, Jonas RA, Castaneda AR, Farrell DM, Lock JE, Wessel DL. Factors associated with choreoathetosis after cardiopulmonary bypass in children with congenital heart disease. Circulation. 1992;86(suppl II):II-118-II-126.
Ferry PC. Neurologic sequelae of open-heart surgery in children. AJDC. 1990;144:369-373.
Kirklin JW, Barratt-Boyes BG. Hypothermia, circulatory arrest, and cardiopulmonary bypass. In: Kirklin JW, Barratt-Boyes BG, eds. Cardiac Surgery. New York: John Wiley and Sons Inc; 1986:37-38.
Ohtake S, Mault JR, Lilly MK, Lilly RE, Kern FH, Greeley WJ, Ungerleider RM. Effect of a systemic-pulmonary artery shunt on myocardial function and perfusion in a piglet model. Surg Forum. 1991;42:200-203.
Sapirstein LA. Regional blood flow by fractional distribution of indicators. Am J Physiol. 1958;193:161-168.
van der Linden J, Ekroth R, Lincoln C, Pugsley W, Scallan M, Tyden H. Is cerebral blood flow/metabolic mismatch during rewarming a risk factor after profound hypothermic procedures in small children? Eur J Cardiothorac Surg. 1989;3:209-215.