Increased pressure during retrograde cerebral perfusion in an acute porcine model improves brain tissue perfusion without increase in tissue edema

Increased Pressure During Retrograde Cerebral Perfusion in an Acute Porcine Model Improves Brain Tissue Perfusion Without Increase in Tissue Edema Zhijun Li, MD, Luojia Yang, MD, Michael Jackson, PhD, Randy Summers, MS, Maureen Donnelly, BS, Roxanne Deslauriers, PhD, and Jian Ye, MD Department of Biochemistry and Medical Genetics, Physiology, and Surgery, University of Manitoba, and Institute For Biodiagnostics, National Research Council of Canada, Winnepeg, Manitoba, Canada

Background. There is a significant lack of scientific data to support the clinically accepted view that 25 to 30 mm Hg is the maximum safe perfusion pressure during retrograde cerebral perfusion (RCP). This study was designed to investigate whether perfusion pressure greater than 30 mm Hg during RCP is beneficial to the brain during prolonged HCA in an acute porcine model. Methods. Sixteen pigs underwent 120 minutes of circulatory arrest in conjunction with RCP at a perfusion pressure of either 23 to 29 mm Hg (group L, n ⴝ 8) or 34 to 40 mm Hg (group H, n ⴝ 8) at 15°C, followed by 60 minutes of normothermic cardiopulmonary bypass. Cortical blood flow and oxygenation were measured continuously with a laser flowmeter and near-infrared spectroscopy, respectively. Tissue water content was measured at the end of the experiments. Results. Brain tissue blood flow was significantly higher in group H than in group L (16.8% ⴞ 4.1% vs 4.8% ⴞ 0.9% of baseline, p < 0.01) during RCP. Brain oxygen

extraction in group L reached a maximum (⬃70%) immediately after starting RCP, whereas in group H it increased gradually and reached a maximum at 120 minutes of RCP, indicating a greater supply of oxygen to tissue in group H than in group L. After RCP, the ability of brain tissue to use oxygen was better preserved in group H than in group L, as indicated by tissue oxygen saturation and the deoxyhemoglobin level. There was no significant increase in tissue water content in either group (group H 79.2% ⴞ 0.3%, group L 79.1% ⴞ 0.4%) relative to normal control pigs (78.7% ⴞ 0.1%). Conclusions. In this acute porcine model, increasing perfusion pressure from 23–29 to 34 – 40 mm Hg during RCP increases tissue blood flow and provides better tissue oxygenation, without increasing tissue edema. The optimal perfusion pressure for RCP needs to be further investigated. (Ann Thorac Surg 2002;73:1514 –21) © 2002 by The Society of Thoracic Surgeons

R

human subjects [4 –7] have failed to demonstrate that RCP provides a significant amount of blood flow to the brain at a perfusion pressure of less than 25 mm Hg or that RCP is able to maintain brain metabolism during HCA. The key to effective brain protection during RCP may therefore lie in the choice of perfusion pressure. Although 25 to 30 mm Hg has been presumed to be the maximum perfusion pressure, there is a significant lack of scientific data to support this concept. Some clinical reports show that RCP, at relatively high pressures (⬎25 mm Hg) and flows, leads to good clinical outcomes with no clinical evidence of either cerebral edema or hemorrhage [8]. Clearly the question of maximum safe perfusion pressure during RCP requires further study. In the present study cerebral blood flow during RCP was monitored by laser Doppler flowmetry, a standard technique for flow measurements. Cerebral oxygenation was monitored using an experimental technique, near-infrared spectroscopy. Near-infrared (NIR) spectroscopy provides a sensitive means of detecting ischemic tissue by measuring

etrograde cerebral perfusion (RCP) has been used for about 10 years as an adjunct to hypothermic circulatory arrest (HCA) in the repair of aortic dissection and aneurysmal disease requiring an open aortic arch. The theoretical advantages of the approach [1] include delivery of oxygen and metabolic substrates to the brain, removal of products of metabolism, and better pH control by the removal of lactic acid and air as well as maintenance of more uniform cerebral temperature during circulatory arrest. Although some clinical reports [2, 3] have suggested that RCP in combination with HCA decreases the rate of stroke and operative mortality associated with aortic arch operation, reports on the efficacy of RCP for brain protection are inconsistent. Experimental and clinical studies have not consistently demonstrated that RCP provides meaningful nutrient flow to support cerebral metabolism during HCA. For example, studies in various experimental models and in Accepted for publication Feb 2, 2002. Address reprint requests to Dr Ye, Institute for Biodiagnostics, 435 Ellice Avenue, Winnepeg, Manitoba, Canada R3B 1Y6; e-mail: [email protected].

© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc

0003-4975/02/$22.00 PII S0003-4975(02)03497-5

Ann Thorac Surg 2002;73:1514 –21

LI ET AL PRESSURE IN RETROGRADE CEREBRAL PERFUSION

1515

Table 1. Experimental Protocol Action

CPB

CPB (Cooling)

CPB

RCP

R-CPB

R-PB (Warning)

R-CPB

37 20

37315 30 –35

15 20

15 120

15 20

15337 35– 40

37 60

Temperature (°C) Duration (min)

CPB ⫽ cardiopulmonary bypass;

R-CPB ⫽ reperfusion with cardiopulmonary bypass;

the chemical composition of tissues. Materials such as lipids, proteins, water, oxyhemoglobin, deoxyhemoglobin (reduced hemoglobin that does not contain oxygen), and cytochrome aa3 each absorb characteristic wavelengths of infrared light. By analyzing the relative proportions of light absorbed at each wavelength, a chemical fingerprint of tissue can be obtained. For instance, a relatively strong absorption feature at 500 to 600 nm arises from hemoglobin species and can provide information relating to the oxygenation status of tissues. Further information can be obtained from analysis of a weak absorption feature at 760 nm that arises from deoxyhemoglobin and a broad absorption feature at 900 nm that is attributed to oxyhemoglobin. Tissue water content can be assessed by analysis of a number of strong infrared absorption bands arising from water, most notably the absorption band at 960 nm [9, 10]. The combination of NIR spectroscopy, laser Doppler flowmetery, and standard laboratory techniques (wet/dry weight, O2 tension) can provide information on the effects of relatively high RCP pressures on cerebral blood flow, oxygenation, tissue water content, and changes in brain oxygen consumption during profound HCA. The present study aimed to determine whether 25 to 30 mm Hg is the maximum safe, effective pressure for RCP during HCA in an acute porcine model.

Material and Methods Sixteen pigs aged less than 5 months and weighing 56 to 73 kg were used, after an acclimatization period of at least 12 days in the animal facility of the Institute for Biodiagnostics. All pigs were fasted with access to water for 12 hours before surgery. All animals received humane care in compliance with the guidelines of the Canadian Council on Animal Care.

Surgical Preparation As described previously [5, 6, 11, 12], preanesthesia was induced with midazolam (0.3 mg 䡠 kg⫺1 IM), ketamine (20 mg 䡠 kg⫺1 IM), and atropine (0.02 mg 䡠 kg⫺1). Muscle relaxation was obtained with pancuronium 0.1 mg 䡠 kg⫺1. After endotracheal intubation, the pig was ventilated mechanically with 60% oxygen and 40% air. The ventilator rate and tidal volume were adjusted to maintain arterial CO2 tension between 35 and 45 mm Hg. Anesthesia was maintained with 1.5% to 2.0% isoflurane. Urine was collected through a bladder catheter. A temperature probe was placed in the esophagus. Because direct monitoring of brain temperature is not practical

RCP ⫽ retrograde cerebral perfusion.

and is not used clinically, esophageal temperature was used to monitor body core temperature. The right temporal muscle was exposed, retracted, and partially excised. Two small holes (0.4 cm and 0.2 cm in diameter) were made in the skull bone using a burr drill. The dura was exposed and remained intact. The holes were prepared for placement of NIR spectroscopy and laser flowmeter probes. A median sternotomy was used to expose the heart. A small catheter was placed in the brachiocephalic artery through the right internal mammary artery for measuring blood pressure and taking venous blood samples during RCP. Another small catheter was placed through the left internal mammary vein, into the right internal jugular vein beyond a venous valve to measure perfusion pressure during RCP and for blood sampling. After heparinization with 500 IU 䡠 kg⫺1, the cardiopulmonary bypass (CPB) circuit was set up with cannulation of the ascending aorta (22F cannula) and the right atrium (28F single-stage venous cannula). The superior vena cava (SVC) was cannulated with a modified 2-lumen cannula. The large lumen (2-mm ID) was used for RCP through the SVC and the small lumen was used to monitor central venous pressure. The lungs were not inflated during CPB or circulatory arrest. The CPB circuit consisted of Cobe roller pumps (model c22.2, Cobe, Arvada, CO), cardiotomy reservoir (Cobe HVRF 3700), arterial filter (40 ␮m, dideco D733, Mirandola, Italy), water bath (Lauda MGW type RMSG, Postfach, Germany), and membrane oxygenator (Cobe Optima) with integrated heat exchanger. The system was primed with 1,000 mL lactated Ringer’s solution, 500 mL Pentaspan, 25 mL of 1 mol/L sodium bicarbonate, and 5000 IU heparin. Sodium bicarbonate was administered as needed to maintain arterial blood pH within the normal range of 7.35 to 7.45. The ␣-stat strategy of acid– base management was used during hypothermia. The CPB circuit was designed to allow switching between RCP and CPB.

Experimental Groups and Protocol Sixteen pigs were randomly assigned to one of the following two groups. Group L (n ⫽ 8) received deep HCA plus RCP at a perfusion pressure of 23 to 29 (average of 27) mm Hg, whereas group H (n ⫽ 8) received deep HCA plus RCP at a perfusion pressure of 34 to 40 (average of 36) mm Hg. The experimental protocol is shown in Table 1. After setting up equipment for NIR spectroscopy and the laser flowmeter (see below), normothermic CPB

1516

LI ET AL PRESSURE IN RETROGRADE CEREBRAL PERFUSION

(37°C) was initiated and continued for 15 to 20 minutes to allow stabilization of body temperature and blood gases. During CPB, pump flow was maintained at 69 to 76 mL 䡠 kg⫺1 䡠 min⫺1 with a mean blood pressure of 58 to 70 mm Hg. After obtaining baseline values for all variables, the pig was gradually cooled from 37° to 15°C with a temperature gradient of less than 10°C between the water bath and blood, which took 30 to 35 minutes. Circulatory arrest was achieved when the esophageal temperature reached 15°C. RCP was performed for 120 minutes through the catheter in the superior vena cava. During RCP, deoxygenated blood returned to the cardiotomy reservoir through the aortic cannula and gravity drainage of the inferior vena cava (IVC). The azygos vein remained patent. The RCP perfusion pressure measured in the internal jugular vein was continuously monitored and carefully controlled within the range specified for each group. At the end of circulatory arrest, CPB was resumed and the pig was gradually rewarmed to 37°C with a temperature gradient of less than 10°C between the water bath and blood. Normothermic CPB was continued for an additional 60 minutes. No inotropic drugs were used during the experiments. Immediately after each experiment, the brain was removed for measurement of brain tissue water content and histopathology.

Measurements The theory and use of the laser flowmeter has been described in detail elsewhere [13, 14]. Regional cerebral blood flow (rCBF) was continuously monitored using an ALF 21R (Advance Company Ltd, Tokyo, Japan) laser Doppler flowmeter fitted with a needle-type probe (Type Nspi:9051U) mounted on a homemade holder. The probe was carefully advanced to touch the dura without visibly indenting it. Areas with visible large blood vessels were avoided. The data were recorded in absolute blood flow units (mL 䡠 100 g⫺1 䡠 min⫺1) once the reading was stable.

CEREBRAL CORTICAL BLOOD FLOW.

Cerebral oxygenation was monitored using NIR spectroscopy. Near-infrared spectra were acquired using a Foss Analytical NIR Systems 6500 NIR spectrometer equipped with a randomized bifurcated fiber optic bundle. The end of the fiber optic probe was positioned on the cerebral temporal cortex of the right side of the brain through a small hole (0.4-cm diameter) in the skull and was held in place with sutures and a homemade holder. An area without any major visible vessels was used for acquisition of the NIR spectra. For each measurement, 32 scans were acquired and summed to produce spectra. Three baseline spectra were acquired during normothermic CPB, with spectra acquired every 5 minutes during the remainder of the experimental protocol.

CORTICAL TISSUE OXYGENATION.

BRAIN OXYGEN CONSUMPTION. Arterial and venous blood samples were obtained simultaneously at each stage of the protocol to monitor blood gases, pH, and electrolytes. Venous or deoxygenated blood samples were collected from the right internal jugular vein and common carotid

Ann Thorac Surg 2002;73:1514 –21

artery during CPB and RCP, respectively. Blood gases were measured immediately after sample collection using a blood gas analyzer (Stat 9, NOVA Biomedical, Waltham, MA). An ␣-stat strategy (measured values were not temperature-corrected to the pig’s actual body temperature) was used to manage blood pH. Oxygen content was calculated based on the formula: O2 content (vol %) ⫽ [HB] ⫻ 1.36 ⫻ SO2 ⫹ Po2 ⫻ 0.003 (HB ⫽ hemoglobin concentration, SO2 ⫽ oxygen saturation). Oxygen extraction was calculated using the formula: oxygen extraction (%) ⫽ (inflow O2 content – outflow O2 content)/inflow O2 content ⫻ 100%. Blocks of tissue from different regions in the brains were obtained and weighed. After weighing, the tissue was dried at 60°C and weighed daily until a constant weight was obtained (typically 72 hours). Tissue water content (%) ⫽ (wet weight – dry weight)/ (wet weight) ⫻100%.

BRAIN TISSUE WATER CONTENT.

HISTOPATHOLOGIC EXAMINATION. The experimental details of the histopathologic studies are described in a previous publication [6]. At the end of each experiment, the brain was removed and fixed with formaldehyde solution. Tissue samples obtained from seven different regions were cut into 5-␮m-thick slices, which were stained with hematoxylin and eosin. Injury was graded on a scale of 0 to 5 based on the number of damaged neurons within each region, as follows: grade 0 ⫽ normal; grade 1 ⫽ less than 10%; grade 2 ⫽ 10% to 25%; grade 3 ⫽ 26% to 50%; grade 4 ⫽ 51% to 75%; and grade 5, more than 75%.

Mean cerebral blood flow, oxyhemoglobin, deoxyhemoglobin, and total brain water signal obtained during initial normothermic CPB were used as baseline levels and set at 100%. Statistical analysis was performed using the Statistical Analysis System (SAS Institute, Cary, NC). All data are presented as mean ⫾ standard error of the mean (SEM). A repeated-measures analysis of variance and Duncan’s multiple range test were used for comparison between different time points within a group, and Students’ t test was used for comparison between the two groups. A p value of less than 0.05 was considered significant.

STATISTICAL ANALYSIS.

Results Animal Characteristics The mean weight of the pigs was 61.6 ⫾ 1.1 kg and 61.3 ⫾ 1.4 kg in groups L and H, respectively, with no significant differences between the two groups. As shown in Table 2, hemodynamics were stable during the entire protocol in both groups. There were no differences between the two groups in blood pressure or pump flow during CPB. Hemoglobin levels in both groups decreased slightly during cooling due to fluid retention in the body requiring the use of additional crystalloid solution during cooling. There were no significant differences in the change in hemoglobin between the two groups. Blood CO2 and pH were maintained within normal ranges

Ann Thorac Surg 2002;73:1514 –21

LI ET AL PRESSURE IN RETROGRADE CEREBRAL PERFUSION

1517

Table 2. Blood Gases, Hematocrit, Mean Arterial Pressure, Esophageal Temperature, and Pump Flow Variable Group pH L H PaCO2 (mm Hg) L H Hematocrit (%) L H Mean BP (mm Hg) L H Esophageal temperature (°C) L H Pump flow (mL 䡠 kg⫺1 䡠 min⫺1) L H

37°C CPB (at 20 min)

15°C CPB (at 20 min)

15°C RCP (at 120 min)

15°C R-CPB (at 20 min)

37°C R-CPB (at 30 min)

37°C R-CPB (at 60 min)

7.39 ⫾ 0.01 7.43 ⫾ 0.01

7.43 ⫾ 0.02 7.45 ⫾ 0.01

7.41 ⫾ 0.04 7.45 ⫾ 0.03

7.31 ⫾ 0.02 7.28 ⫾ 0.04

7.42 ⫾ 0.04 7.44 ⫾ 0.02

7.32 ⫾ 0.02 7.38 ⫾ 0.01

47.7 ⫾ 2.05 44.7 ⫾ 2.23

34.5 ⫾ 2.39 35.4 ⫾ 1.43

38.7 ⫾ 4.73 43.0 ⫾ 3.85

43.7 ⫾ 2.04 47.6 ⫾ 2.85

36.9 ⫾ 2.45 34.7 ⫾ 1.57

41.8 ⫾ 2.12 39.2 ⫾ 1.18

24.7 ⫾ 0.56 24.2 ⫾ 1.11

17.7 ⫾ 0.95 18.5 ⫾ 0.76

16.7 ⫾ 0.67 17.3 ⫾ 0.84

19.7 ⫾ 0.67 17.3 ⫾ 0.84

19.0 ⫾ 0.77 20.8 ⫾ 0.54

18.8 ⫾ 1.19 19.3 ⫾ 1.08

62.1 ⫾ 1.55 63.8 ⫾ 3.93

61.0 ⫾ 3.98 61.0 ⫾ 3.71

27.0 ⫾ 0.52 36.0 ⫾ 0.63a

58.3 ⫾ 1.15 58.5 ⫾ 2.22

63.3 ⫾ 3.44 60.3 ⫾ 2.59

59.3 ⫾ 4.07 60.5 ⫾ 3.01

37.2 ⫾ 0.12 37.2 ⫾ 0.06

14.5 ⫾ 0.14 14.4 ⫾ 0.16

14.5 ⫾ 0.24 14.4 ⫾ 0.18

15.3 ⫾ 0.13 15.3 ⫾ 0.11

37.1 ⫾ 0.08 37.1 ⫾ 0.06

37.0 ⫾ 0.11 37.2 ⫾ 0.11

72.9 ⫾ 1.53 74.9 ⫾ 4.67

69.7 ⫾ 1.62 72.4 ⫾ 4.80

5.0 ⫾ 0.18 13.8 ⫾ 2.64a

69.4 ⫾ 2.77 73.8 ⫾ 4.17

73.8 ⫾ 3.14 75.6 ⫾ 4.24

73.7 ⫾ 2.17 75.9 ⫾ 5.21

Values are mean ⫾ SEM. a

p ⬍ .05 vs group L.

CPB ⫽ cardiopulmonary bypass; RCP ⫽ retrograde cerebral perfusion; at 23–29 mm Hg; R-CPB ⫽ reperfusion with CPB.

H ⫽ group receiving RCP at 34 – 40 mm Hg;

throughout the experimental protocol in both groups. There was no difference in esophageal temperature between the two groups.

pressure of 34 to 40 mm Hg (average 36 mm Hg) in group H. There was significant statistical difference in the pump/retrograde flow during RCP between the two groups (p ⬍ 0.01).

Pump/Retrograde Flow During RCP As shown in Figure 1, pump flow during normal CPB was 70 to 75 mL 䡠 kg⫺1 䡠 min⫺1 in both groups. During RCP, the pump/retrograde flow was set at 3.8 to 5.0 mL 䡠 kg⫺1 䡠 min⫺1 to achieve an internal jugular vein pressure of 23 to 29 mm Hg (average 27 mm Hg) in group L, and 13.5 to 14.1 mL 䡠 kg⫺1 䡠 min⫺1 to achieve a internal jugular vein

L ⫽ group receiving RCP

Regional Cerebral Blood Flow Regional cerebral blood flow (rCBF) was measured continuously throughout the experimental protocol. The mean value obtained during normothermic CPB was initially used as the baseline (100%). Regional cerebral blood flow decreased to 48% of the baseline level during CPB at 15°C, indicating a significant effect of temperature on tissue blood flow. Regional cerebral blood flow returned to baseline after rewarming in both groups (Fig 2A). To eliminate the effect of temperature, rCBF measured during CPB at 15°C was used as the baseline (100%) for assessment of tissue blood flow during RCP (Fig 2B). Regional cerebral blood flow during RCP was significantly higher (p ⬍ 0.01) in the high perfusion pressure group (11.4% to 16.8% of baseline flow) than in the low perfusion pressure group (4.3% to 4.6% of baseline flow).

Tissue Oxygenation Fig 1. Pump flow during experiments in the low-pressure (designated in text as L; squares) and high-pressure (designated in text as H; triangles) groups. (CPB ⫽ cardiopulmonary bypass; CPB-b ⫽ baseline levels obtained during initial CPB; RCP ⫽ retrograde cerebral perfusion; Temp. ⫽ temperature.) (*p ⬍ 0.05 vs low-pressure group.)

Figure 3 shows four typical in vivo NIR spectra obtained from a pig brain at four different stages of the experimental protocol. From these spectra, the peak area of deoxyhemoglobin, oxyhemoglobin, and water were quantitated using a computer program [9, 10]. Oxyhemoglobin and deoxyhemoglobin were expressed as the ratio of

1518

LI ET AL PRESSURE IN RETROGRADE CEREBRAL PERFUSION

oxyhemoglobin/total hemoglobin and deoxyhemoglobin/total hemoglobin. The level obtained during initial normothermic CPB was used as the baseline level (100%). The oxyhemoglobin level significantly increased when body temperature reached 15°C during CPB and significantly decreased during RCP in both groups. There were no statistical differences in the changes in oxyhemoglobin during 15°C CPB and RCP between groups L and H. After rewarming, oxyhemoglobin levels in group H were slightly lower than the baseline level and significantly lower than the levels in group L, in which the oxyhemoglobin level was greater than the baseline level (Fig 4A). In contrast, the deoxyhemoglobin level decreased during hypothermic CPB and increased during RCP in both groups (218.8% ⫾ 46.3% of baseline in group L, and 180.6% ⫾ 32.7% in group H at 120 minutes of RCP). The changes during hypothermic CPB and RCP were similar in groups L and H. After rewarming, the deoxyhemoglobin level in group H returned to baseline and was significantly higher than that in group L (107.7% ⫾ 12.7% vs 61.6% ⫾ 6.5% of the baseline 60 minutes after rewarming, p ⬍ 0.01). The deoxyhemoglobin level did not return to baseline in group L (60% of the baseline level) (Fig 4B).

Ann Thorac Surg 2002;73:1514 –21

Fig 3. Representative in vivo near-infrared spectra obtained from a pig brain during initial cardiopulmonary bypass (CPB) (baseline), at 60 minutes of retrograde cerebral perfusion (RCP), at 120 minutes of RCP, and during reperfusion with CPB at 15°C. The figure depicts a relatively strong absorption feature at 500 to 600 nm that arises from hemoglobin (oxyhemoglobin and deoxyhemoglobin) to become a single peak as deoxyhemoglobin increases; a weak absorption feature at 760 nm that arises from deoxyhemoglobin; a broad absorption feature at 900 nm that is attributed to oxyhemoglobin; and an absorption band at 960 nm that arises from water.

Tissue oxygen saturation was calculated using the formula: tissue oxygen saturation (%) ⫽ oxyhemoglobin/ (oxyhemoglobin ⫹ deoxyhemoglobin) ⫻ 100%. Tissue oxygen saturation during initial normothermic CPB was 86.6% ⫾ 2.1% and 86.2% ⫾ 2.9% in groups L and H, respectively. The changes in tissue oxygen saturation during hypothermic CPB and RCP were similar in both groups. Tissue oxygen saturation in group H returned to baseline and was significantly lower than that in group L after rewarming after RCP (85.7% ⫾ 2.8% of base line in group H vs 92.5% ⫾ 1.2% of baseline in group L 60 minutes after rewarming, p ⬍ 0.05) (Fig 4C). This indicates that the ability of brain tissue to use oxygen remained at its baseline level in the high–pressure group and decreased in the low–pressure group after 120 minutes of RCP.

Brain Oxygen Extraction

Fig 2. Change in brain tissue blood flow determined by laser flowmetry during experiments in both the low-pressure (designated in text as L; squares) and high-pressure (H; triangles) groups. (A) Levels obtained during initial normothermic cardiopulmonary bypass (CPB) were used as baseline (CPB-b, 100%). (B) Levels obtained during initial CPB at 15°C were used as baseline (CPB-b, 100%). (RCP ⫽ retrograde cerebral perfusion; Temp. ⫽ temperature.) (*p ⬍ 0.05 vs low-pressure group.)

Brain oxygen extraction during initial normothermic CPB was 23.1% ⫾ 2.7% and 22.7% ⫾ 4.7% in groups L and H, respectively. When switching to RCP, brain oxygen extraction increased immediately to near maximum in group L (62.5% ⫾ 3.4%). In group H, brain oxygen extraction initially increased to 51.9% ⫾ 5.5% and then gradually increased to its maximum level (66.4% ⫾ 3%) at 120 minutes of RCP. This suggests that more oxygen was provided to the brain during RCP at high perfusion pressure, however, there was no statistically significant difference between two groups. After rewarming, brain oxygen extraction returned to baseline in both groups (Fig 5).

Ann Thorac Surg 2002;73:1514 –21

LI ET AL PRESSURE IN RETROGRADE CEREBRAL PERFUSION

1519

Fig 5. Brain oxygen extraction during experiments in both the lowpressure (designated in text as L; squares) and high-pressure (designated in text as H; triangles) groups. The asterisks and plus sign indicate p ⬍ 0.05 vs levels obtained at 15 and 30 minutes of retrograde cerebral perfusion (RCP) within the group. There were no differences between the low- and high-pressure groups. (CPB ⫽ cardiopulmonary bypass; CPB-b ⫽ baseline levels obtained during initial CPB; Temp. ⫽ temperature.)

experimental group L (79.09% ⫾ 0.38%) or group H (79.16% ⫾ 0.34%) relative to control pigs. There was no difference in tissue water content between groups L and H.

Histopathology

Fig 4. Changes in brain tissue oxyhemoglobin (A) and deoxyhemoglobin (B), as well as brain tissue oxygen saturation (C) determined by near-infrared spectroscopy during experiments in both the lowpressure (designated in text as L; squares) and high-pressure (designated in text as H; triangles) groups. In (A) and (B), the levels obtained during initial normothermic cardiopulmonary bypass (CPB) were used as baselines (CPB-b, 100%). (RCP ⫽ retrograde cerebral perfusion; Temp. ⫽ temperature.) (*p ⬍ 0.05 vs low-pressure group.)

Neuronal injury was observed in pigs that received 120 minutes of RCP at either low or high perfusion pressures. The mean grade was 1.12 in the low-pressure group, and 1.10 in the high-pressure group. No significant difference was observed between the two groups.

Changes in Total Water Signal The signal arising from intracellular, interstitial, and intravascular water was determined by NIR spectroscopy. The water signal decreased significantly during RCP in both groups. The decrease was greater in group L than in group H (82.4% ⫾ 4.2% of baseline vs 91.9% ⫾ 3.6% of baseline at 120 minutes of RCP), but this difference was not statistically significant. The decrease in water signal was most likely due to lower tissue blood flow in group L. After rewarming, the water signal returned to its baseline level in both groups (Fig 6).

Brain Tissue Water Content Final brain tissue water content measured by the wet– dry method was 78.70% ⫾ 0.15% in normal control pigs. There was no increase in tissue water content in either

Fig 6. Change in total brain tissue water (including intracellular, interstitial, and intravascular water) determined by near-infrared spectroscopy during experiments in the low-pressure (designated in text as L; squares) and high-pressure (designated in text as H; triangles) groups. The levels obtained during initial normothermic cardiopulmonary bypass (CPB) were used as baseline values (CPB-b, 100%). (RCP ⫽ retrograde cerebral perfusion; Temp. ⫽ temperature.) There were no differences between the low- and high-pressure groups.

1520

LI ET AL PRESSURE IN RETROGRADE CEREBRAL PERFUSION

Comment Retrograde cerebral perfusion has been considered to provide metabolic substrates and oxygen to the brain during HCA and is widely used in clinical practice. The maximum pressure for RCP has been presumed to be 25 to 30 mm Hg, although there is little clinical or experimental evidence to support this concept. As a general consideration, under physiologic conditions, capillary pressures beyond 25 mm Hg can cause tissue edema. However, during RCP perfusion pressure is measured in large veins and does not represent the pressure in capillary systems. Pressure in the capillaries is believed to be significantly lower than that in large veins during RCP. Therefore, a retrograde perfusion pressure greater than 25 mm Hg may not cause tissue edema, and some clinical reports [8] suggest that RCP at relatively high perfusion pressures (⬎ 25 mm Hg) appears to be safe, with good clinical outcomes and no clinical evidence of either cerebral edema or hemorrhage. Few animal studies [15, 16] exist to validate the use of a maximal 25 to 30 mm Hg perfusion pressure. We are not aware of any other animal or clinical studies on the relationship between RCP pressure and brain tissue blood flow/edema that have been reported supporting the concept that 25 to 30 mm Hg is the maximum safe perfusion pressure for RCP. Studies by others and ourselves using pigs, baboons, sheep, and human cadavers have demonstrated that RCP at perfusion pressures less than 25 to 30 mm Hg provides very limited blood flow to brain tissue and minimal [4, 7, 11, 12, 17] or no brain protection [18]. Although cerebral blood flow is one of the most important factors in ensuring adequate brain protection, accurate monitoring of cerebral blood flow during RCP is still a challenge. Noninvasive techniques are extremely important for accurate measurement of cerebral blood flow because local tissue hemorrhage/injury occurs even with the placement of a fine needle, as observed in our other experiments. Currently, no single method is capable of monitoring tissue blood flow accurately in the entire brain during RCP. Tracking injected microspheres is considered to be the “gold standard” in many situations. However, inconsistencies have been reported with this method, which may be related to precipitation of microspheres in large-diameter veins with low flow. For continuous noninvasive monitoring of regional blood flow, laser Doppler flowmetry can be used by careful placement of the probe on the surface of the dura. Although measurement of absolute tissue blood flow with laser Doppler flowmetry is of limited accuracy, the method can monitor relative changes in regional tissue blood flow under varying conditions in vivo [13]. Laser Doppler flowmetry has also been used clinically to measure cerebral blood flow during RCP [3]. There are no established techniques for measuring cerebral oxygenation. Near-infrared spectroscopy has been used in both humans [19, 20] and animals in experimental situations. The technique allows noninvasive, almost real-time measurements of local tissue oxyhemoglobin, deoxyhemoglobin, and tissue oxygen satu-

Ann Thorac Surg 2002;73:1514 –21

ration in vivo. In our studies, the NIR probe was placed directly against the dura to eliminate signals from the skull and skin and to allow accurate monitoring of changes in cortical tissue oxygenation. We found that tissue oxygen saturation returned to baseline levels during reperfusion after 120 minutes of RCP at perfusion pressures of 34 to 40 mm Hg, but was higher than baseline in pigs that received RCP at perfusion pressures of 23 to 29 mm Hg. Brain tissue blood flow was similar in both groups during reperfusion. This may indicate that increased retrograde perfusion pressure (34 to 40 mm Hg) better preserves the ability of neurons to use oxygen. The ability of neurons to use oxygen during reperfusion in the low-pressure group appeared to be reduced, which may be a consequence of neuron injury or “neuronal stunning.” Histopathologic studies did not show any differences between the two groups. This may be because (1) microscopic morphology is not sufficiently sensitive to detect small changes that occur in an acute study; (2) functional changes occur before morphologic changes; and (3) delayed neuronal injury cannot be observed in the acute model. The study clearly supports the view that increasing perfusion pressure from 23 to 29 to 34 to 40 mm Hg during RCP significantly increases brain tissue blood flow (in this study by a factor of 3), and improves oxygen supply to brain tissue (as indicated by the slow increase in brain oxygen extraction during RCP at higher pressure), leading to better preservation of the ability of brain tissue to use oxygen during reperfusion. More importantly, RCP at pressures of 34 to 40 mm Hg did not cause brain tissue edema in this acute model, as determined by NIR spectroscopy and tissue water content. The venous system behaves as a very large sump such that the increase in venous pressure for perfusion was only 5 to 17 mm Hg while the pump/retrograde flow was tripled. Our results are in contrast to those of Nojima and colleagues [16] and Usui and associates [15] in a canine model. This may be explained by the several factors. First, different animal models were used: Nojima and colleagues and Usui and associates used a canine model. Second, in their studies, RCP was established through the maxillary veins with temporary occlusion of the SVC, IVC and azygos vein, which is similar to total body retrograde perfusion. In our study, the brain was perfused retrogradely through the SVC without occlusion of the IVC and azygos vein. Different management of the IVC and the azygos vein may play a major role in development of brain edema and injury during RCP [8, 21–23]. Third, an invasive method (a probe placed in brain tissue) was used by Nojima and colleagues and by Usui and associates for the measurement of cerebral blood flow. In our studies, the laser Doppler flowmeter probe did not penetrate the brain tissue, and did not cause any brain tissue injury. Fourth, different sites were used for measurements of perfusion pressure. Both groups used the external jugular venous pressure as the retrograde perfusion pressure, whereas the internal jugular venous pressure was considered to be the retrograde perfusion pressure in our study.

Ann Thorac Surg 2002;73:1514 –21

Because of possible anatomical differences between humans and animals, and because we used an acute porcine model, our data cannot be completely translated into clinical situations. However, animal models provide controlled experimental conditions and allow measurements that often are not feasible in humans. The present study provides a detailed report on regional cerebral blood flow, brain tissue oxygenation, and changes in cerebral tissue water during RCP at different perfusion pressures. A chronic animal model would be ideal for evaluation of possible delayed neurologic changes that cannot be determined in an acute study. In conclusion, in an acute porcine model, increasing perfusion pressure from 23 to 29 mm Hg, to 34 to 40 mm Hg during RCP provides more tissue blood flow and better tissue oxygenation, without any increase in tissue edema. A pressure of 25 to 30 mm Hg may not be the maximum safe perfusion pressure for RCP. The optimal perfusion pressure for RCP needs to be further investigated, particularly in a chronic animal model. This work was supported in part by the Canadian Institutes of Health Research (Grant Nos. 15352 and 42671) and the Manitoba Health Research Council through grants. We thank Lori Gregorash, Allan Turner, Rachelle Mariash, and Shelly Germscheid for technical assistance. We also thank Dr Mike Sowa for his assistance in the analysis of NIR spectra.

References 1. Bavaria JE, Pochettino A. Retrograde cerebral perfusion (RCP) in aortic arch surgery: efficacy and possible mechanisms of brain protection. Semin Thorac Cardiovasc Surg 1997;9:222–32. 2. Ueda Y, Miki S, Kusuhara K, Okita Y, Tahata T, Yamanaka K. Deep hypothermic systemic circulatory arrest and continuous retrograde cerebral perfusion for surgery of aortic arch aneurysm. Eur J Cardiothorac Surg 1992;6:36– 41. 3. Lin PJ, Chang CH, Tan PP, et al. Prolonged circulatory arrest in moderate hypothermia with retrograde cerebral perfusion. Is brain ischemic? Circulation 1996;94(Suppl II):II-169 – II-72. 4. Boeckxstaens CJ, Flameng WJ. Retrograde perfusion does not perfuse the brain in non-human primates. Ann Thorac Surg 1995;60:319–28. 5. Filgueiras CL, Ryner L, Ye J, et al. Cerebral protection during moderate hypothermic circulatory arrest: histopathology and magnetic resonance spectroscopy of brain energetics and intracellular pH in pigs. J Thorac Cardiovasc Surg 1996; 112:1073– 80. 6. Ye J, Yang L, Del Bigio MR, et al. Neuronal damage after hypothermic circulatory arrest and retrograde perfusion in the pig. Ann Thorac Surg 1996;61:1316–22.

LI ET AL PRESSURE IN RETROGRADE CEREBRAL PERFUSION

1521

7. Katz MG, Khazin V, Steinmetz A, et al. Distribution of cerebral flow using retrograde versus antegrade cerebral perfusion. Ann Thorac Surg 1999;67:1065–9. 8. Ganzel BL, Edmonds HL Jr, Pank JR, Goldsmith LJ. Neurophysiologic monitoring to assure delivery of retrograde cerebral perfusion. J Thorac Cardiovasc Surg 1997;113: 748–57. 9. Sowa MG, Leonardi L, Matas A, et al. In-vivo tissue analysis by near infrared spectroscopy. In: Meyers RA, ed. Encyclopedia of Analytical Chemistry, vol 1. Chichester: Wiley, 2000:251– 81. 10. Matcher SJ, Elwell CE, Cooper CE, Cope M, Delpy DT. Performance comparison of several published tissue nearinfrared spectroscopy algorithms. Anal Biochem 1995;227: 54– 68. 11. Ye J, Yang L, Del Bigio MR, et al. Retrograde cerebral perfusion provides limited distribution of blood to the brain: a study in pig. J Thorac Cardiovasc Surg 1997;114:660–5. 12. Ye J, Ryner LN, Kozlowski P, et al. Retrograde cerebral perfusion results in flow distribution abnormalities and neuronal damage. A magnetic resonance imaging and histopathological study in pigs. Circulation 1998;98:II-313–II-8. 13. Dirnagl U, Kaplan B, Jacewicz M, Pulsinelli W. Continuous measurement of cerebral cortical blood flow by laserDoppler flowmetry in a rat stroke model. J Cereb Blood Flow Metab 1989;9:589–96. 14. Guo J, Liao JJ, Preston JK, Batjer HH. A canine model of acute hindbrain ischemia and reperfusion. Neurosurgery 1995;36:986–93. 15. Usui A, Oohara K, Liu TL, et al. Determination of optimum retrograde cerebral perfusion conditions. J Thorac Cardiovasc Surg 1994;107:300– 8. 16. Nojima T, Magara T, Nakajima Y, et al. Optimal perfusion pressure for experimental retrograde cerebral perfusion. J Card Surg 1994;9:548–59. 17. Filgueiras CL, Winsborrow B, Ye J, et al. A 31P-magnetic resonance study of antegrade, and retrograde cerebral perfusion during aortic arch surgery in pigs. J Thorac Cardiovasc Surg 1995;110:55– 62. 18. Crittenden MD, Roberts CS, Rosa, et al. Brain protection during circulatory arrest. Ann Thorac Surg 1991;51:942–7. 19. Katoh T, Esato K, Gohra H, et al. Evaluation of brain oxygenation during selective cerebral perfusion by nearinfrared spectroscopy. Ann Thorac Surg 1997;64:432– 6. 20. Rigg CD, Clutton-Brock TH. Near-infrared spectroscopy changes during hypothermic circulatory arrest with retrograde cerebral perfusion. Anaesthesia 1997;52:356–9. 21. Yasuura K, Ogawa Y, Okamoto H, et al. Clinical application of total body retrograde perfusion to operation for aortic dissection. Ann Thorac Surg 1992;53:655– 8. 22. Matalanis G, Buxton BF. Retrograde vital organ perfusion during aortic arch repair. Ann Thorac Surg 1993;56:981– 4. 23. Juvonen T, Zhang N, Wolfe D, et al. Retrograde cerebral perfusion enhances cerebral protection during prolonged hypothermic circulatory arrest: a study in a chronic porcine model. Ann Thorac Surg 1998;66:38–50.