Impact of Pump Flow Rate During Selective Cerebral Perfusion on Cerebral Hemodynamics and Metabolism

Transcription

1 Impact of Pump Flow Rate During Selective Cerebral Perfusion on Cerebral Hemodynamics and Metabolism Peter L. Haldenwang, MD, Justus T. Strauch, MD, Igor Amann, Tobias Klein, Anja Sterner-Kock, PhD, Hildegard Christ, and Thorsten Wahlers, MD Department of Cardiothoracic Surgery, Experimental Medicine, and Institute for Medical Statistics, Informatics and Epidemiology, University of Cologne, Cologne, Germany Background. Although hypothermic selective cerebral perfusion (SCP) is widely used for cerebral protection during aortic surgery, little is known about the ideal pump-flow management during this procedure. This study explored cerebral hemodynamics and metabolism at two different flow rates. Methods. Fourteen pigs (33 to 38 kg) were cooled on cardiopulmonary bypass to 25 C. After 10 minutes of hypothermic circulatory arrest, the animals were randomly assigned to 60 minutes of SCP at two different pump flow rates: 8 ml kg 1 min 1 (n 7) and 18 ml kg 1 min 1 (n 7). Microspheres were injected at baseline, coolest temperature, and at 5, 15, 25, and 60 minutes of SCP to calculate cerebral blood flow, cerebral vascular resistance, metabolic rate, and intracranial pressure. Results. Cerebral blood flow decreased during cooling to 41% of the baseline value (from to 23 4 ml min g 1 ). It recovered during the initial 15 minutes of SCP, showing a significantly higher increase (p 0.017) at high-flow versus low-flow perfusion ( versus ml min g 1 ). After 60 minutes of SCP the cerebral blood flow almost returned to baseline values in the low-flow group (43 25 ml min g 1 ), but showed an unexpected decrease (30 7 ml min g 1 ) in the high-flow group. The highest regional cerebral blood flow was seen in the cortex (66 12 ml min g 1 ), followed by the cerebellum (63 12 ml min g 1 ), the pons (51 17 ml min g 1 ), and the hippocampus (36 9 ml min g 1 ). Intracranial pressure increased from 11 3to13 5mm Hg during cooling on cardiopulmonary bypass. During low-flow SCP, it stayed stable at baseline values, whereas high-flow perfusion resulted in significantly higher intracranial pressures (17 3mmHg;p 0.001). Changes in cerebral vascular resistance and metabolic rate showed no significant differences between the groups. Conclusions. High-flow SCP provides no benefit during long-term SCP at 25 C. Higher cerebral blood flow during the initial SCP period leads to cerebral edema, with no profit in metabolic rate. (Ann Thorac Surg 2010;90: ) 2010 by The Society of Thoracic Surgeons The surgical approach of aortic arch resection is known to be associated with an elevated incidence of neurologic dysfunctions and requires special cerebral protection strategies [1, 2]. The affordance of a bloodless operation field, as well as the requirement of an adequate cerebral blood perfusion according to the metabolic oxygen rate of the brain, has led to the combination of two neuroprotective techniques: a short period of hypothermic circulatory arrest (HCA), ideal for an open inspection of the aortic arch and a safe cannulation of the carotid artery, followed by an extended interval of antegrade selective cerebral perfusion (SCP), enabling the surgeon to carry out complex surgical procedures without strict time limitations, owing to the maintenance of cerebral blood flow (CBF) [3, 4]. There is clinical and experimental evidence of improved neurologic outcome using SCP Accepted for publication June 23, Address correspondence to Dr Haldenwang, Department of Cardiothoracic Surgery, University of Cologne, Kerpener Str. 62, Cologne, 50924, Germany; peter-lukas.haldenwang@uk-koeln.de. rather than HCA alone when prolonged cerebral protection is required [5 7]. However, little is known about the ideal pump flow rate during SCP. Excessively high-flow pump rates may increase vascular resistance and hydrostatic pressure, which in combination with an altered endothelial function during extended cardiopulmonary bypass (CPB) may cause cerebral edema, but these mechanisms of brain injury, as well as the warm reperfusion damage, are not fully understood. Otherwise, a low-flow SCP management may lead to a regional hypoperfusion and ischemia in cerebral arterial border zones [8]. At hypothermic conditions the brain is known to be more tolerant of ischemia; however, there is little evidence about the relationship between the true regional blood flow (RBF) and the degree of metabolism suppression in different brain areas. Crittenden and associates [9] were the first to demonstrate in a sheep model that hypothermic low-flow SCP preserves intracellular ph and energy stores. Less is still known about regional changes in blood flow during SCP at different pump flow levels by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1976 HALDENWANG ET AL Ann Thorac Surg PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 2010;90: Abbreviations and Acronyms CBF cerebral blood flow CPB cardiopulmonary bypass CMRO 2 cerebral metabolic rate of oxygen CVR cerebral vascular resistance HCA hypothermic circulatory arrest ICP intracranial pressure RBF regional blood flow SCP selective cerebral perfusion SSP sagittal sinus pressure We undertook this study in pigs to investigate how a short, clinically relevant interval of HCA followed by an extended SCP period at different perfusion pressure levels affects the RBF, cerebral oxygen metabolism (CMRO 2 ), cerebral vascular resistance (CVR), intracranial pressure (ICP), and sagittal sinus pressure (SSP). Material and Methods Fourteen female juvenile pigs, approximately 3.5 months of age, with a weight of 30 to 38 kg, were used for this experiment. All animals received preoperative humane care in compliance with the guidelines of the North- Rhine-Westphalian Chamber of Agriculture. The protocol for the experiment was approved by the German Research Society (Deutsche Forschungsgesellschaft). Pigs were randomly assigned to one of the following groups: group 1, 10 minutes of HCA followed by 60 minutes of SCP at 8 ml kg 1 min 1 (n 7); or group 2, 10 minutes of HCA followed by 60 minutes of SCP at 18 ml kg 1 min 1 (n 7). Perioperative Management and Anesthesia After pretreatment with intramuscular azaperone (2 mg/kg) and ketamine (15 to 20 mg/kg) to induce deep sedation, animals were anesthetized with intravenous propofol 1% (1 to 2 mg/kg/h), fentanyl (25 g kg 1 h 1 ), and midazolam (0.2 mg kg 1 h 1 ). After endotracheal intubation, the pigs were ventilated mechanically with a fraction of inspired oxygen of 0.5. Paralysis was achieved with intravenous pancuronium (0.2 mg kg 1 h 1 ). The ventilation rate and the tidal volume were adjusted (Fabius, Dräger, Germany) to maintain the arterial carbon dioxide tension at about 35 to 40 mm Hg. Arterial oxygen tension was maintained greater than 100 mm Hg. A bladder catheter (Foley 10F) was inserted for online measurement of urine output, and temperature probes were placed in the rectum and the brain through a small burr hole in the skull. A 14-gauge arterial line was placed in the right brachial artery for pressure monitoring and arterial blood sampling (Blood Gas Analyzer; ABL Radiometer, Copenhagen, Denmark). Intracranial Pressure Sagittal sinus cannulation was performed before cannulation and heparinization for CPB. A midline scalp incision was made, and the underlying periosteum was removed to facilitate identification of the coronal and sagittal sutures. A 5-mm cutting burr was used to remove the bone over the sinus. A 24-gauge catheter was inserted into the sagittal sinus to permit both sampling of cerebral venous blood and monitoring of cerebral venous pressure. The ICP probe was connected to a transducer (Codman ICP Express; Codman, Raynham, MA). Operative Technique The chest was opened by means of a small left thoracotomy in the fourth intercostal space. After opening the pericardium the heart and the great vessels were exposed. The ascending and descending aorta were dissected, and vessel loops were placed to define the future levels of clamping. After heparinization (300 IU/kg) the distal ascending aorta was cannulated with a 16F arterial cannula, and the right atrium, with a single 26F cannula. Nonpulsatile CPB, using alpha-stat ph management, was initiated at a systemic flow rate of 80 to 100 ml kg 1 min 1 and then adjusted to maintain a minimum mean arterial pressure of 50 mm Hg. To avoid distention of the left ventricle during CPB and as an injection port for fluorescent microsphere injection, a 10F vent catheter was inserted through the left atrium. After initiation of CPB the lungs were allowed to collapse. A heat exchanger was used for core cooling. The CPB circuit included roller pumps (Stöckert Instruments, Munich, Germany), a cardiotomy reservoir, and a membrane oxygenator (VPCML Plus, Cobe Cardiovascular Inc, Arvada, CO) that was primed with previous citrate-treated blood from an animal donor, heparin (5000 IU), and KCl (1.5 meq/kg). The ph was maintained, by means of alpha-stat principles, at 7.40 with an arterial Pco 2 of 35 to 40 mm Hg and uncorrected for temperature. Hemoglobin level was maintained between 8 and 10 g/dl. Once stable CPB was established, cooling for approximately 45 to 60 minutes to a brain temperature of 25 C was undertaken (heat exchanger: Biomedicus; Medtronic, Minneapolis, MN), and the operating room temperature was maintained at 22 C to prevent an upward temperature drift. Cardiac arrest was achieved by clamping the ascending aorta and adding blood cardioplegic solution (blood plus 30 ml of 14.9% KCl plus 10 ml of 50% magnesium) with an infusion rate of 3.5 ml/min for 2 minutes (induction) and then 2 ml/min for 2 minutes (maintenance, every 20 minutes during cross-clamp time) through a 4F cannula into the ascending aorta. Additional myocardial protection was afforded by applying cold saline solution (approximately 8 C) topically in the pericardium. Selective cerebral perfusion was installed after 10 minutes of HCA by clamping the descending aorta. Isolation of the aortic arch permitted SCP through the brachiocephalic trunk and the left subclavian artery. In this concept the subclavian arteries were included in the SCP circuit for arterial pressure monitoring (upper limb) and reference blood sample withdrawal for microsphere flow calculation (Fig 1).

3 Ann Thorac Surg HALDENWANG ET AL 2010;90: PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 1977 Fig 1. Diagram showing cardiopulmonary bypass assessment. (Ao aorta; LV left ventricle; LV vent left ventricular vent cannulation; Ox oxygenator; P arterial pressure measurement and reference microsphere-withdrawal; PO microsphere injection; RA right atrium; Res reservoir; SCP selective cerebral perfusion.) After 10 minutes of HCA and 60 minutes of SCP, CPB was reinstituted by releasing the clamp in the descending aorta. Core and surface rewarming were begun and continued to a brain temperature of approximately 35 to 36 C. Care was taken to avoid a temperature difference between the perfusate and core temperature of greater than 10 C. During weaning from CPB, administration of 3to5mg kg 1 min 1 dobutamine was frequently used. When necessary, cardiac defibrillation was performed after administration of lidocaine (1 mg/kg). Cerebral Blood Flow Cerebral blood flow was measured with fluorescent microspheres as described in previous studies [10]. In brief, approximately 2 million microspheres, m in diameter, in six different colors were injected and flushed with 5 ml of saline solution into a left ventricular catheter before CPB and into the aortic cannula during SCP. Before injection, the fluorescently labeled microspheres, suspended in 10% dextran with 0.05% polyoxyethylene sorbitan monooleate (Tween 80), were mixed, sonicated, and vortexed. To allow calculation of absolute blood flow rates, a reference blood sample was taken from the brachial artery (upper right limb) at a rate of 2.9 ml/min with a Harvard withdrawal pump (Harvard Bioscience, Inc, Holliston, MA). The withdrawal of blood started 10 seconds before injection of the microspheres and continued for 110 seconds after the microsphere injection. Microspheres of 15 m were selected to avoid failure to be entrapped, a risk when using smaller microspheres, and to avoid streaming and hemodynamic sequelae from occlusion of larger than capillary vessels, a risk when using larger microspheres [11]. The animals were sacrificed 60 minutes after aortic cannula removal with an intravenous injection of sodium pentobarbital (30 mg/kg) and saturated potassium chloride (6 meq/kg). In all animals the brain was removed, the two hemispheres cut in the middle, and the specimens were weighed. Tissue samples (0.6 to 1.2 g) from five different regions (frontal and parietal neocortex, cerebellum, hippocampus, and brainstem) were taken for microsphere count. Thereafter, the microspheres were recovered from the brain tissue by sedimentation and from the blood by using a commercial protocol (NuFlow Extraction protocol ; Interactive Medical Technologies Ltd, Irvine, CA). Fluorescent analysis was carried out by following a protocol from the same company. Regional CBF was then calculated from the intensity of fluorescence microspheres in blood and tissue samples using the following formula: CBF ml min g 1 R I T / I R Wt where R was the rate at which the reference blood sample was withdrawn (2.9 ml/min), I T was the fluorescence intensity of the tissue sample, I R was the fluorescence intensity of the blood sample, and Wt was the weight of the tissue sample (in grams). Cerebral Metabolism Cerebral sagittal sinus and arterial samples were obtained simultaneously for calculation of cerebral oxygen extraction (arteriovenous oxygen content difference), sagittal sinus oxygen saturation, and cerebral oxygen saturation extraction (arteriovenous oxygen saturation difference). Cerebral vascular resistance was calculated by using the following equation, where MAP refers to mean arterial pressure and MSSP is mean SSP: CVR MAP MSSP /CBF Cerebral metabolic rate of oxygen was determined as follows: CMRO 2 ml min g 1 CBF arterial O 2 content sagittal sinus O 2 content 100 Arterial and venous blood ph, oxygen tension, carbon dioxide tension, hemoglobin level, oxygen saturation, and oxygen content as well as glucose and lactate in the sagittal sinus were measured using the same blood gas analyzer (ABL Radiometer). Study Protocol Hemodynamics such as heart rate, central venous pressure, mean arterial pressure, and SSP as well as core and brain temperature were monitored continuously (Omni-

4 1978 HALDENWANG ET AL Ann Thorac Surg PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 2010;90: care 24C; Hewlett Packard, Böblingen, Germany). Additionally, ICP, arterial and sagittal sinus blood gases, hemoglobin, glucose, and lactate were recorded at six times as shown in Figure 2. Regional and total cerebral CBF, CVR, and CMRO 2 were calculated at the same times: at baseline, after reaching the temperature of 25 C (coolest T C), and at 5, 15, 25, and 60 minutes of SCP. Statistical Methods Randomization was carried out by an independent party (cardioperfusion technician), with individual group allocation revealed at the onset of SCP. The distribution of values was assured by a Shapiro-Wilk test. When the data were consistent with normality and equal variance assumptions, the groups were compared at different times using repeated measures analysis of variance, with tests for average differences between groups and for time-by-group interactions (change in the difference between groups as a function of time). Otherwise the groups were compared separately at each time using the Mann-Whitney U test (between groups testing). Pairwise comparisons between groups (time analyses) were conducted if the corresponding average difference or time-by-group interaction was significant. Hotelling-Spur and Friedman s test were used, as appropriate, for analyzing changes as a function of time for one group (within group testing). We report probability values unadjusted for multiple testing: their purpose is not meant as an exact global assessment but rather as a guide to help interpret the pattern of differences between groups at different times. Analyses were implemented by a certified biostatistician with SPSS software (SPSS, Inc, Chicago, IL). Results Comparability of Experimental Groups All animals reached the final measurement time of 30 minutes after removal from CPB. A comparison of preoperative animal weights (group 1, kg versus group 2, kg) showed no differences between the groups. As intended by the design of the study, blood gas analysis showed no significantly relevant differences between groups in arterial ph, partial oxygen and carbon dioxide pressures, and hemoglobin values (Table 1). The same was the case with the brain temperatures (Table 2), heart rate, and central venous pressure. Significantly higher mean arterial pressure values were seen in the high-flow group during the entire SCP period (p 0.001; Table 2). Global and Regional Cerebral Blood Flow GLOBAL CEREBRAL BLOOD FLOW. The values for the global CBF at the six different times for each group are shown in Table 3. There were no significant differences for CBF at baseline. Cerebral blood flow decreased to 41% of the baseline value during cooling to 25 C on CPB, showing no differences between the groups. During the initial period of SCP, a rapid return to normal values could be seen in group 1, even exceeding the baseline after 15 and 25 minutes of SCP (125% and 147% of baseline, respectively). In group 2, the initial 15 minutes of SCP were associated with a dramatic increase of CBF (259% of baseline), followed by a substantial decrease in CBF during the last 35 minutes of SCP, finally reaching lower global CBF values than in group 1. Significant differences between groups were seen at 5 and 15 minutes of SCP (p 0.004, p 0.017, respectively; deviation from baseline, p 0.004, p 0.007, respectively). REGIONAL BLOOD FLOW. At baseline the highest RBF was seen in the frontal cortex ( ml min g 1 ) and the parietal cortex ( ml min g 1 ), followed by the cerebellum ( ml min g 1 ) and the pons ( ml min g 1 ). The lowest RBF at baseline was measured in the hippocampus ( ml min g 1 ). The general trend of change in global CBF with time was reproducible among all regions of the brain. The detailed RBF values for the five separate brain regions at the different times are presented in Table 4. The RBF time course is shown Figure 3. FRONTAL AND PARIETAL CORTEX. In the frontal and parietal cortex, cooling to 25 C reduced the RBF to 34% to 42% of the initial baseline value. After installing SCP, the RBF Fig 2. Study protocol, showing times at which measurements cited in the tables and figures were taken. Details are listed in the text. The two groups differ beginning with the third period (5 min SCP). (CPB cardiopulmonary bypass; HCA hypothermic circulatory arrest; MAP mean arterial pressure; SCP selective cerebral perfusion; T temperature.)

5 Ann Thorac Surg HALDENWANG ET AL 2010;90: PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 1979 Table 1. Results of Blood Gas Analysis a Variable Baseline Coolest Temperature C 5 min SCP 15 min SCP 25 min SCP 60 min SCP p Value b pha Group 1 (low-flow) Group 2 (high-flow) Hb (g/dl) Group 1 (low-flow) Group 2 (high-flow) Pao 2 (mm Hg) Group 1 (low-flow) Group 2 (high-flow) Paco 2 (mm Hg) Group 1 (low-flow) Group 2 (high-flow) a All values are shown as mean standard deviation. b Between group: p values for two-way analysis of variance; within group: changes as a function of time within one group (Hotelling-Spur or Friedman s test). Hb hemoglobin; Pao 2 alveolar oxygen pressure; Paco 2 partial pressure of carbon dioxide, alveolar; pha arterial ph; SCP selective cerebral perfusion. increased in both groups during the first 15 minutes, with a higher RBF in the frontal and parietal cortex at 5 minutes of SCP in group 2 (132% of baseline). The cortical RBF continued to increase in the low-flow group, reaching a maximal value slightly higher than the baseline at 25 minutes of SCP, followed by a decrease to 45% (frontal) and 57% (parietal) of the initial baseline value after 60 minutes of SCP. The cortical RBF showed an unexpected decrease during the last 45 minutes of SCP in the high-flow group, with a final value of ml min g 1 (35% of baseline) in the frontal and mL min g 1 (44% of baseline) in the parietal cortex both being lower than the final RBF values in the low-flow group. No significant differences in cortical RBF were seen at any time between the two groups. CEREBELLUM. The RBF of the cerebellum also decreased during cooling, reaching 40% of baseline at 25 C. It returned and remained stable at physiologic values during the first 25 minutes of low-flow SCP (group 1), showing afterward a decrease to 59% of the baseline during the last 35 minutes of low-flow SCP. In group 2, the RBF increased dramatically during the initial period of high-flow SCP, reaching 230% of the baseline value after 15 minutes, but presented a more substantial decrease during the last 45 minutes of SCP compared with the low-flow group: at 60 minutes of SCP, the RBF of the cerebellum dropped to 46% of the baseline value (28.9 Table 2. Temperature and Hemodynamic Data a Variable Baseline Coolest Temperature ( C) 5 min SCP 15 min SCP 25 min SCP 60 min SCP p Value b Brain temperature ( C) Group 1 (low-flow) Group 2 (high-flow) MAP (mm Hg) Group 1 (low-flow) Group 2 (high-flow) Time point analysis p value b ICP (mm Hg) Group 1 (low-flow) Group 2 (high-flow) Time point analysis p value b SSP (mm Hg) Group 1 (low-flow) Group 2 (high-flow) Time point analysis p value b a All values are shown as mean standard deviation. b Between group: p values for two-way analysis of variance; Time point analysis was done if p 0.05 in analysis of variance testing; Within group: changes as a function of time within one group (Hotelling-Spur or Friedman s test). ICP mean intracranial pressure; MAP mean arterial pressure; SCP selective cerebral perfusion; SSP mean sagittal sinus pressure.

6 1980 HALDENWANG ET AL Ann Thorac Surg PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 2010;90: Table 3. Global Cerebral Blood Flow, Vascular Resistance, Oxygen Extraction, and Metabolic Rate of Oxygen a Variable Baseline Coolest Temperature C 5 min SCP 15 min SCP 25 min SCP 60 min SCP p Value b CBF (ml min g 1 ) Group 1 (low-flow) Group 2 (high-flow) Time point analysis p value b CVR (mm Hg ml min 100 g) Group 1 (low-flow) Group 2 (high-flow) O 2 Extraction (ml/dl) Group 1 (low-flow) Group 2 (high-flow) CMRO 2 (ml min g 1 ) Group 1 (low-flow) Group 2 (high-flow) a All values are shown as mean standard deviation. b Between group: p values for two-way analysis of variance; time point analysis was done if p 0.05 in analysis of variance testing; within group: changes as a function of time within one group (Friedman s test). CBF mean global cerebral blood flow; CMRO 2 mean cerebral metabolic rate of oxygen; CVR mean cerebral vascular resistance; O 2 Extraction oxygen content difference between arterial and sagittal sinus blood; SCP selective cerebral perfusion. 8 ml min g 1 ) in the high-flow group. The differences between groups were statistically significant at 5 and 15 minutes of SCP (p 0.038, p 0.007, respectively) and in changes from baseline at 15 minutes (p 0.011). PONS. During cooling to 25 C, the RBF of the pons decreased to 53% of the baseline value. At SCP it rapidly returned to baseline values after 5 minutes, continuing to increase moderately during the first 25 minutes in the low-flow group, but showed a dramatic increase to 390% of baseline after 5 minutes and to 500% of baseline after 15 minutes of high-flow SCP. These differences between groups were significant in absolute values (p at 5 minutes, p at 15 Table 4. Regional Blood Flow in Different Brain Areas a Brain Area RBF (ml min g 1 ) Baseline Coolest Temperature C 5 min SCP 15 min SCP 25 min SCP 60 min SCP p Value b Frontal cortex Group 1 (low-flow) Group 2 (high-flow) Parietal cortex Group 1 (low-flow) Group 2 (high-flow) Cerebellum Group 1 (low-flow) Group 2 (high-flow) Time point analysis p value b Pons Group 1 (low-flow) Group 2 (high-flow) Time point analysis p value b Hippocampus Group 1 (low-flow) Group 2 (high-flow) Time point analysis p value b a All values are shown as mean standard deviation. b Between group: p values for two-way analysis of variance; time point analysis was done if p 0.05 in analysis of variance testing; within group: changes as a function of time in regional blood flow for one group (Hotelling-Spur or Friedman s test). RBF regional blood flow; SCP selective cerebral perfusion.

7 Ann Thorac Surg HALDENWANG ET AL 2010;90: PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 1981 Fig 3. Time course of regional cerebral blood flow (RBF) in five brain areas: frontal and parietal cortex (A), cerebellum (B), pons (C), and hippocampus (D). Values are shown as mean values, and * indicates significant differences between groups. (SCP selective cerebral perfusion; T temperature.) minutes) as well as in deviation from baseline (p at 5 minutes, p at 15 minutes). The RBF decreased in both groups during the last 35 minutes of SCP, reaching values slightly higher than the baseline in the low-flow group (70 38 ml min g 1 ), but showing a lower value than the normal RBF rate in the high-flow group (39 11 ml min g 1 ). HIPPOCAMPUS. The RBF changes of the hippocampal area showed a similar pattern as in the cerebellum: after a decrease at cooling, the RBF returned to close to baseline values and remained stable at normal values during the entire SCP period in the low-flow group. At high-flow perfusion (group 2), the RBF greatly exceeded the normal values during the first 25 minutes of SCP, reaching a peak at 15 minutes of SCP with 400% of the baseline value, but decreased during the last 35 minutes of SCP to normal RBF values. Statistically, the RBF values both in absolute terms and as changes from baseline differed significantly at 5 and 15 minutes of SCP (absolute values: p 0.002, p 0.002, respectively; change from baseline: p 0.007, p 0.002, respectively). Cerebral Vascular Resistance Cerebral vascular resistance (Table 3) did not show significant differences between the groups at baseline ( versus mm Hg ml 1 min 100 g), increasing in all animals during cooling to 25 C. After starting SCP, the CVR stayed stable in both groups at 0.6 to 0.8 mm Hg ml 1 min 100 g during the first 25 minutes, but increased during the last 35 minutes of SCP to mm Hg ml 1 min 100 g in group 1 and mm Hg ml 1 min 100 g in group 2. Although both groups showed an increase in CVR during the last 35 minutes of SCP, there were no significant differences between absolute values and changes from baseline as a function of time. Cerebral Oxygen Metabolism The CMRO 2 presented a significant decrease from mean baseline values of to ml min g 1 (29% of the baseline value, p in group 1, p in group 2) in all animals during cooling to 25 C. After starting the SCP, the CMRO 2 increased to 51% of the baseline in group 1 and 75% in group 2, followed by a second drop, decreasing slightly during the remaining SCP period to 38% of the baseline in group 1 and 25% in group 2 (Table 3). Values of CMRO 2 as their deviation from baseline showed no significant differences between groups at any time. Intracranial Pressure At baseline the ICP showed no significant differences between groups ( mm Hg versus mm Hg). During cooling to 25 C it stayed almost stable at baseline values, slightly increasing in the high-flow group ( mm Hg). Differences between groups could be seen after starting SCP at different pump-flow rates: whereas the ICP decreased slowly to 9.4 2mm Hg, staying stable at 82% to 96% of the baseline for the entire SCP period in the low-flow group, it showed a sudden rise to 135% of the baseline value during the first 5 minutes of high-flow SCP, continuing to increase to mm Hg for the rest of the SCP interval (Table 2, Fig 4). The measurements at 5, 15, 25, and 60 minutes of SCP revealed significant differences between groups in absolute ICP values (p at 5 minutes, p at 25 minutes, p at 60 minutes) as well as in the ICP deviation from baseline (p at 5 minutes, p at 15 minutes, p at 25 minutes, p at 60 minutes).

8 1982 HALDENWANG ET AL Ann Thorac Surg PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 2010;90: Fig 4. Time course of intracerebral pressure (ICP) during the entire experiment. Values are shown as mean values standard error, and * indicates significant differences between groups. The absolute ICP values did not differ at 15 min SCP (p 0.084). Only the changes from each baseline differ significantly (p 0.026). (SCP selective cerebral perfusion; T temperature.) Sagittal Sinus Pressure The SSP changes as a function of time presented a similar pattern as the ICP values did (Fig 5). At baseline (9.4 2 mm Hg versus mm Hg) and during cooling (7.7 3 mm Hg versus mm Hg), no differences between groups could be seen. Whereas low-flow perfusion resulted in a further decrease in SSP to 71% to 86% of the baseline value in group 1, high-flow perfusion induced a sudden increase in SSP up to mm Hg after 5 minutes and mm Hg after 60 minutes of SCP in group 2 (Table 2). The groups showed significant differences at 5, 15, 25, and 60 minutes of SCP in absolute SSP values (p 0.005, p 0.011, p 0.015, p 0.001, respectively) as well as in the SSP change from baseline (p 0.007, p 0.019, p 0.030, p 0.002, respectively). Comment Hypothermic circulatory arrest and antegrade SCP represent frequently used neuroprotective procedures during aortic arch surgery. Nevertheless, the ideal strategy in terms of cooling temperature, safe duration, and adequate pump flow rate during SCP remains a major topic in neuroprotective research. Traditionally HCA and SCP were used at deep hypothermic temperatures (temperature 22 C) to offer a maximum of cerebral metabolic suppression [3, 12]. More recent experimental and clinical studies have shown that HCA and SCP at milder hypothermic conditions (temperature of 25 C) provide similar neuroprotective results for a duration of 60 minutes [13], and that shorter periods of SCP (25 minutes) could be performed in 120 consecutive patients at a mean core temperature of 30 C with an excellent neurologic outcome [14]. This new trend toward more tepid temperatures with the advantage of shorter operating times and lower bleeding tendency also found application in our experimental study. In the literature there is still no consensus regarding the optimal pump flow rate during SCP: each institution has its own method to perform SCP. In general a systemic pump flow rate of 100 ml kg 1 min 1 is used for total body perfusion during CPB. When changing to SCP, the pump flow rate is reduced to 10 to 30 ml kg 1 min 1, representing 10% to 30% of the systemic value [15]. During SCP, a pump flow rate of 10 ml kg 1 min 1 in an adult of 70 to 80 kg results in a blood flow of 700 to 800 ml/min into the upper body, irrigating the brain and part of the upper extremities, with an almost negligible quantity being shunted away into the lower torso. If we consider a flow of 700 ml/min reaching the brain with an average weight of 1300 to 1400 g, it comes to a perfusion rate of 50 to 54 ml min g 1, close to the physiologic CBF in humans of ml min g 1 [16]. Nevertheless, there is a concern that these values considered low-flow values may lead to regional hypoperfusion owing to intracerebral shunt mechanisms. Otherwise it has to be considered that autoregulation mechanisms of the cerebral vessels may be altered under hypothermic conditions: a too-high pump flow can produce luxury perfusion with local edema in some regions of the brain, whereas other cerebral areas with an increased vascular resistance may suffer a regional ischemia. In our porcine model, we found some differences between the brain areas: at baseline the highest RBF was seen in the neocortex and cerebellum, followed by the pons region and finally the hippocampal area. All cerebral regions followed the same pattern of RBF change as a function of time, with a constant increase during the first 15 minutes of SCP (high-flow) or 25 minutes of SCP (low-flow) and a decrease during the last 45 minutes of SCP (high-flow) or 35 minutes of SCP (low-flow). After 60 minutes of SCP, the RBF returned to normal values in the areas with the physiologic lower CBF (pons and hippocampus), but dropped to lower values than the initial baseline in the neocortex and the cerebellum. An interesting finding was the fact that RBF was higher at 18 ml kg 1 min 1 than at 8 ml kg 1 min 1 only during the initial 15 to 25 minutes of SCP, dropping finally in all examined brain areas to lower levels in the high-flowgroup. The drop of CBF was associated with a tremen- Fig 5. Time course of sagittal sinus pressure (SSP) during the entire experiment. Values are shown as mean values standard error, and * indicates significant differences between groups. (SCP selective cerebral perfusion; T temperature.)

9 Ann Thorac Surg HALDENWANG ET AL 2010;90: PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 1983 dous increase of the ICP ( 16 mm Hg) after 25 minutes of SCP in the high-flow-group. The increase of ICP and SSP during the entire high-flow SCP period could be an explanation of the reduction of global CBF as a result of ongoing cerebral edema under high-flow conditions. The mechanisms for reduced CBF after a prolonged period of SCP, even at high-flow rates, as well as the development of a cerebral edema beyond the 25 minutes benchmark are not totally clear. It seems likely that the high hydrostatic pressure in combination with a variable colloid osmotic pressure and stimulation of acute phase reactants owing to a prolonged CPB time are responsible for a substantial capillary leak syndrome. This is also accentuated by the ischemic membrane damage during the HCA period. The leakage of fluid into the extravascular space raises the ICP, compressing the small cerebral vessels and lowering the CBF. In a previous study, we showed in a chronic porcine model that high ICP may cause depression of CBF and ultimately, if severe, destruction of brain tissue [10]. Comparing our metabolic data with the reference literature, we achieved somewhat higher CMRO 2 baseline values ( ml/min/100 g) than the group of Ehrlich and associates [17] with ml min g 1 in 7- to 13-kg pigs. If we consider the human CMRO 2 varying between 2.55 and 3.45 ml min g 1 [18, 19] both being calculated according to according to the Kety-Schmidt technique [18] our findings in 33- to 38-kg pigs come pretty close to the adult human values. A concordance was seen also during cooling on CPB: the pigs lowered their metabolic rate to ml min g 1 at 25 C, whereas humans showed in a similar protocol a reduction to ml min g 1 by cooling to 27 C. Both porcine study groups showed an initial increase of cerebral metabolism during the first 5 minutes of SCP. This is not surprising after a total ischemia of 10 minutes, even at 25 C. Nevertheless, this rise did not exceed 75% of the normal CMRO 2. Although the relative metabolic suppression was more pronounced in the high-flow group, the low-flow perfusion offered a sufficient neuroprotection because the CMRO 2 was even lower than at baseline. One of the limitations of our experimental setup is the perfusion of both the internal as well as the external carotid arteries. Thus, the SCP includes external perfusion of the porcine skull and neck muscles. Because we know that the porcine brain is smaller, but the neck musculature more pronounced than in humans, this may lead to a different distribution of blood flow during SCP. Nevertheless this experimental setup is a well-established model for SCP monitoring also used by other groups [17, 20], being the study model mostly comparable to the clinical situation. In conclusion, the results of this study indicate that high-flow SCP (18 ml kg 1 min 1 ) performed for 60 minutes at 25 C does not improve regional and global CBF when compared with low-flow perfusion: after 25 minutes it increases both the CVR and the ICP, without showing a significant benefit in CMRO 2 suppression. Low-flow SCP at 8 ml kg 1 min 1 results in lower CBF during the initial 5 minutes and after 25 minutes of SCP when compared with the baseline. However, the achieved CBF seems to be sufficient to protect the brain under moderate hypothermic conditions without increasing the vascular resistance and producing a cerebral edema. References 1. Kazui T, Kimura N, Yamada O, Komatsu S. Surgical outcome of aortic aneurysm using selective cerebral perfusion. Ann Thorac Surg 1994;57: Bachet J, Guilmet D, Goudot B, et al. Antegrade selective cerebral perfusion with cold blood: a 13-year experience. Ann Thorac Surg 1999;67: Ergin MA, Galla JD, Lansman SL, Quintana C, Bodian C, Griepp RB. Hypothermic circulatory arrest on operations on the thoracic aorta: determinants of operative mortality and neurologic outcome. J Thorac Cardiovasc Surg 1994; 107: Strauch JT, Spielvogel D, Haldenwang PL, et al. Cerebral physiology and outcome after hypothermic circulatory arrest followed by selective cerebral perfusion. Ann Thorac Surg 2003;76: Di Eusanio M, Schepens MA, Morshuis WJ, et al. Brain protection using selective antegrade selective perfusion: a multicenter study. Ann Thorac Surg 2003;76: Fraser CD Jr, Andropoulos DB. Principles of antegrade cerebral perfusion during arch reconstruction in newborns/ infants. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2008;11: Strauch JT, Spielvogel D, Lauten A, et al. Technical advances in total aortic arch replacement. Ann Thorac Surg 2004;77: Howard R, Trend P, Russell RW. Clinical features of ischemia in cerebral arterial border zones after periods of reduced cerebral blood flow. Arch Neurol 1987;44: Crittenden MD, Roberts CS, Rosa L, et al. Brain protection during circulatory arrest. Ann Thorac Surg 1991;51: Strauch JT, Spielvogel D, Haldenwang PL, et al. Impact of hypothermic selective cerebral perfusion compared with hypothermic cardiopulmonary bypass on cerebral hemodynamics and metabolism. Eur J Cardiothorac Surg 2003;24: Fox LS, Blackstone EH, Kirklin JW, Bishop SP, Bergdahl LAL, Bradley EL. Relationship of brain blood flow and oxygen consumption to perfusion flow rate during profoundly hypothermic cardiopulmonary bypass. J Thorac Cardiovasc Surg 1984;87: Griepp RB, Stinson EB, Hollingsworth JF, Buehler D. Prosthetic replacement of aortic arch. J Thorac Cardiovasc Surg 1975;70: Salazar J, Coleman R, Griffith S, et al. Brain preservation with selective cerebral perfusion for operations requiring circulatory arrest: protection at 25 C is similar to 18 C with shorter operating times. Eur J Cardiothorac Surg 2009;36: Bakhtiary F, Dogan S, Zierer A, et al. Antegrade cerebral perfusion for acute type A aortic dissection in 120 consecutive patients. Ann Thorac Surg 2008;85: Tanaka H, Kazui T, Sato H, Inoue N, Yamada O, Komatsu S. Experimental study on the optimum flow rate and pressure for selective cerebral perfusion. Ann Thorac Surg 1995;59: Lassen NA. Normal average value of cerebral blood flow in younger adults is 50 ml/100 g/min. J Cereb Blood Flow Metab 1985;5: Ehrlich M, McCullough J, Zhang N, et al. Effect of hypothermia on cerebral blood flow and metabolism in the pig. Ann Thorac Surg 2002;73: Cook DJ, Anderson RE, Michenfelder JD, et al. Cerebral blood flow during cardiac operations: comparison of Kety-

10 1984 HALDENWANG ET AL Ann Thorac Surg PUMP FLOW RATE AND SELECTIVE CEREBRAL PERFUSION 2010;90: Schmidt and xenon-133 clearance methods. Ann Thorac Surg 1995;59: Yamaguchi T, Kanno I, Uemura K, et al. Reduction in cerebral metabolic rate of oxygen during human aging. Stroke 1986;17: INVITED COMMENTARY The article by Haldenwang and colleagues [1] from the University of Cologne adds meaningfully to the continuously developing literature on brain protection during circulatory arrest. Although deep hypothermic circulatory arrest is a highly specific part of cardiac surgical practice, it remains an area of extraordinary surgical, anesthetic, and conceptual challenges. This area is uniquely difficult because deep hypothermia and circulatory arrest lie outside the boundaries of the mammalian physiologic and evolutionary experience. The level of complexity is extraordinary. The surgical challenges of aortic arch surgery are overlaid with physiologic considerations of temperature, flow, pressure, carbon dioxide, and hematocrit, and further, with changes in regulation of cerebral blood flow and metabolism. Open surgical repair of the aortic arch and concurrent brain protection techniques are in evolution. In a short span of time, the practice has moved from clamp and go to antegrade carotid perfusion to retrograde venous perfusion to selective antegrade cerebral perfusion by axillary cannulation. In contrast to other surgical areas where technique and management have largely matured, deep hypothermic circulatory arrest and adjunctive protection techniques remain exploratory, and we are learning as we go. This is why well designed and conducted physiologic studies in large animals, such as the one by Haldenwang and colleagues, can contribute meaningfully to practice evolution. The study differs from clinical practice in a few ways. The cerebral circulation of the pig has real anatomic differences from humans, and the technique of cerebral perfusion in this study, perfusion by way of the arch with proximal and distal clamping, does not mimic either what is possible during open repair or the increasingly 20. Hagl C, Khaladj N, Peterss S, et al. Hypothermic circulatory arrest with and without cold selective antegrade cerebral perfusion: impact on neurological recovery and tissue metabolism in an acute porcine model. Eur J Cardiothorac Surg 2004;26: adopted practice of selective antegrade perfusion through the right axillary artery. Nevertheless, the physiologic effect of high-flow, high-pressure perfusion on global and regional cerebral perfusion, cerebral vascular resistance, intracranial, and sagittal sinus pressure remain valid and of clinical significance. The observation that cerebral perfusion is not improved and that intracranial pressure and brain edema worsens in response to high-flow, high-pressure antegrade perfusion is valuable for at least three reasons: first, it is of direct value to understand the specific physiology of this form of brain protection; second, it helps define the boundaries for flow management in clinical practice; and third, it demonstrates that the brain is vulnerable to the circulatory conditions we may create and that its ability to protect itself under such conditions can be overwhelmed, so caution is indicated. David J. Cook, MD Cardiovascular and Thoracic Anesthesiology Division Department of Anesthesiology College of Medicine Mayo Clinic 200 First St SW Rochester, MN cook.david@mayo.edu Reference 1. Haldenwang PL, Strauch JT, Amann I, et al. Impact of pump flow rate during selective cerebral perfusion on cerebral hemodynamics and metabolism. Ann Thorac Surg 2010;90: by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur