Retrograde Cerebral Perfusion Versus Selective Cerebral Perfusion as Evaluated by Cerebral Oxygen Saturation During Aortic Arch Reconstruction

Transcription

1 Retrograde Perfusion Versus Selective Perfusion as Evaluated by Oxygen Saturation During Aortic Arch Reconstruction Tetsuya Higami, MD, Syuichi Kozawa, MD, Tatsuro Asada, MD, Hidefumi Obo, MD, Kunio Gan, MD, Kazuhiko Iwahashi, MD, and Hideaki Nohara, MD Division of Cardiovascular Surgery, Hyogo Brain and Heart Center, Himeji, Japan Background. Time limits for neuroprotection by retrograde cerebral perfusion (RCP) and selective cerebral perfusion (SCP) in aortic arch aneurysm repair or dissection are undergoing definition. Methods. Using near-infrared optical spectroscopy, changes in regional cerebrovascular oxygen saturation (rso 2 ) were compared between the two perfusion methods. Results. Immediately before cardiopulmonary bypass, baseline rso 2 was 63.9% 6.9% for the RCP and 66.1% 5.3% for the SCP group (no significant difference). As patients were core-cooled to 20 C, rso 2 increased to 73.1% 8.8% and 74.1% 7.9% in the RCP and SCP groups, respectively. With circulatory arrest, rso 2 suddenly decreased. After starting cerebral perfusion, rso 2 returned to prearrest values in the SCP group but continued decreasing steadily in the RCP group, to levels below baseline after about 25 minutes. At the end of perfusion, rso 2 was 57.4% 12.2% for the RCP group and 71.7% 6.9% for the SCP group, and the ratio of rso 2 to baseline value was 0.89 for RCP and 1.08 for SCP despite a shorter brain perfusion time for RCP ( versus minutes). Three of 5 patients whose ratios of rso 2 to baseline at the end of brain protection were 0.7 or less had neurologic deficits. Conclusions. Although SCP showed no clinically important time limitation, rso 2 continued to decrease with time during RCP. An rso 2 ratio less than 0.7 could represent a critical lower limit. (Ann Thorac Surg 1999;67:1091 6) 1999 by The Society of Thoracic Surgeons Brain damage is a frequent complication in patients who have aortic arch operations, especially procedures that require circulatory arrest. Deep hypothermic circulatory arrest has been used either alone or in combination with various partial cerebral perfusion techniques for intraoperative protection of the central nervous system. However, neurologic injury remains a serious consequence despite the protective effect of hypothermia [1]. Antegrade and retrograde cerebral perfusion during hypothermic circulatory arrest has been reported to provide better brain protection during operation than hypothermic circulatory arrest alone [2 7]. However, the efficacy of these techniques remains to be fully determined, especially when used for prolonged periods. Moreover, especially for retrograde cerebral perfusion, individual differences at the point of distribution of blood flow to the brain are so great that the maximum safe time should be estimated for each patient [8]. Near-infrared spectroscopy can measure cerebral oxygen saturation only in a limited region and cannot Accepted for publication Oct 13, Address reprint requests to Dr Higami, Division of Cardiovascular Surgery, Hyogo Brain and Heart Center, 520 Saisyo-ko, Himeji, , Japan. monitor the entire brain. In contrast to this limitation of the method are its advantages for intraoperative monitoring, including noninvasiveness and real-time operation [9, 10]. If we can determine the critical limit of cerebral oxygen saturation to prevent ischemic brain damage, cerebral oxygen saturation by near-infrared spectroscopy could be an important a real-time indicator for brain protection. To test this hypothesis, we sought to determine a safe time range for each cerebral perfusion method in terms of brain tissue oxygenation by using near-infrared spectroscopy. We evaluated relationships between cerebral perfusion time, neurologic deficits, and changes in brain tissue oxygenation by using nearinfrared spectroscopy in patients who had antegrade or retrograde cerebral perfusion under hypothermia or hypothermic circulatory arrest, to determine both a critical limit of cerebral oxygen saturation to prevent ischemic brain damage and a safe time range for each cerebral perfusion method. Patients and Methods We reviewed the hospital records and postoperative clinical charts of 92 consecutive patients who had nonemergent surgical treatment for aneurysm or dissection 1999 by The Society of Thoracic Surgeons /99/$20.00 Published by Elsevier Science Inc PII S (99)

2 1092 HIGAMI ET AL Ann Thorac Surg RETROGRADE PERFUSION AND BRAIN OXYGEN 1999;67: involving the aortic arch at our institution between March 1991 and February The study was approved by the Hyogo Brain and Heart Center Institutional Review Board. Regional cerebrovascular oxygen saturation (rso 2 ) was measured in all patients. Brain protection involved retrograde cerebral perfusion (RCP) with profound hypothermia in 40 patients and selective cerebral perfusion (SCP) in 52 patients. Perfusion and Surgical Procedures Retrograde cerebral perfusion through the superior vena cava with hypothermic circulatory arrest [5] was used for proximal aortic arch aneurysms involving the ascending aorta or for aortic dissection where the point of entry was located in the aortic arch, in a total of 15 patients. Median sternotomy was done. After systemic heparinization, the femoral artery, ascending aorta, or both were cannulated for arterial return, and single cannulas were inserted into the superior and inferior vena cavae for venous access. Cardiopulmonary bypass was established, and the patient was core-cooled until the rectal temperature was 20 C. If the ascending aorta could be cannulated, the femoral arterial catheter was clamped during corecooling and used for backflushing of thrombus or air during circulatory arrest. The left ventricle was vented through the right superior pulmonary vein. The superior vena cava was occluded with a snare, and systemic circulatory arrest was induced. The aortic arch was opened with no clamping, and RCP was initiated with oxygenated blood at 16 to 18 C through a cannula inserted into the superior vena cava, with jugular venous pressure maintained at 20 mm Hg at a flow rate of 200 to 350 ml/m 2 per minute. To protect the myocardium, cold blood cardioplegic solution was perfused continuously in a retrograde direction using a Retro-TH catheter (Fuji Systems, Tokyo, Japan) placed in the coronary sinus. Distal anastomosis of a prosthetic graft with the descending aorta or aortic arch was done, and the arch vessels were reconstructed if necessary. Air was removed from the graft, which then was occluded. Perfusion through the graft to the distal aorta was reinstituted to resume cerebral circulation, and rewarming was started. During this time the proximal aorta was anastomosed with the graft. After venting the prosthetic graft, the graft occlusion was released and cardiac resuscitation was commenced by conventional extracorporeal perfusion through the graft or the ascending aortic return alone. The femoral arterial line was clamped to eliminate backflushing of emboli. Retrograde cerebral perfusion by means of femoral arterial perfusion, with patients in the Trendelenburg position [6], was used for distal aortic arch aneurysms in 25 other patients. Left lateral thoracotomy was done. The femoral artery was cannulated, and a drainage cannula was inserted from the femoral vein into the right atrium. A femorofemoral bypass was initiated using a roller pump for arterial return and a centrifugal pump to augment venous drainage without heart venting. If problems were associated with femoral perfusion, a perfusion cannula was inserted in a distal portion of the descending aorta. The patient was cooled to a rectal temperature of 20 C and then was placed in the Trendelenburg position, at a central venous pressure of 20 mm Hg. The descending aorta was clamped proximal to the points of cannulation, and the distal aortic arch aneurysm was incised. The high central venous pressure prevailing in the Trendelenburg position with the aortic arch open allowed oxygen-saturated venous blood to perfuse the brain in a retrograde direction from the right atrium to the carotid artery. A catheter that occludes aortic irrigation was inserted into the ascending aorta, and cardioplegic solution was infused through it. The temperature of blood perfused through the femoral artery ranged from 16 to 18 C. During RCP, proximal anastomosis of the graft with the distal aortic arch was done. The graft was constructed with a side branch through which the brain and heart were reperfused, and rewarming was carried out. During rewarming, distal anastomosis with the descending aorta was done. The heart then was defibrillated, and the side branch of the graft was cut short and sutured except when required for reconstruction of the left subclavian artery. Selective cerebral perfusion using a balloon catheter with no clamp was used for aortic arch aneurysms involving the proximal or distal arch when total reconstruction of arch vessels was required. The technique for SCP was devised by us, involves no clamping, and is referred to as nonclamping selective cerebral perfusion [4]. It uses three balloon catheters to deliver blood to each of the three arch vessels. Median sternotomy was done. After systemic heparinization, the femoral artery, ascending aorta, or both were cannulated for arterial return, and the superior and inferior vena cavae each received a single cannula for venous access. Cardiopulmonary bypass was established, and the patient was core-cooled. If the ascending aorta could be cannulated, the femoral arterial catheter was clamped during corecooling. The left ventricle was vented through the right superior pulmonary vein after ventricular fibrillation occurred with core-cooling. When the tympanic membrane temperature (as opposed to the rectal temperature) reached 20 C, systemic circulatory arrest was induced temporarily. The aortic arch was opened with no clamping, and three SP Stud catheters (Fuji Systems, Tokyo, Japan) were inserted into the three arch vessels (14F for the innominate artery and 12F for the left carotid and subclavian arteries). The SP Stud catheter, which we developed, has a balloon with a ribbed surface to minimize slippage and therefore eliminate any need for snaring or clamping. The catheter has two lumens, for perfusion and for pressure monitoring. Selective cerebral perfusion was established immediately through the catheters with oxygenated blood at 16 to 18 C at a total flow rate of 300 to 350 ml/m 2 per minute, using a single roller pump separated from the systemic circulation to maintain a perfusion pressure of more than 30 mm Hg. At the same time, the descending aorta was blocked by an occlusive balloon catheter inserted through the femoral artery, and distal aortic systemic perfusion was started by way of the cannula in the femoral artery, with a rectal

3 Ann Thorac Surg HIGAMI ET AL 1999;67: RETROGRADE PERFUSION AND BRAIN OXYGEN 1093 Table 1. Profiles of Patients With Retrograde Perfusion and With Selective Perfusion Characteristic Retrograde Perfusion (n 40) Selective Perfusion (n 52) p Value Age (y) NS Sex (M:F) 30:10 35:17 NS True aneurysm 18 (45.0%) 31 (59.6%) NS Dissection 20 (50.0%) 21 (40.4%) NS Reconstruction site Ascending arch 15 (37.5%) 19 (36.5%) NS Arch 0 (0.0%) 8 (15.4%) 0.05 Distal arch 24 (60.0%) 25 (48.1%) NS Concomitant operation Coronary artery bypass 2 (5.0%) 14 (26.9%) 0.01 grafting Aortic valve replacement 1 (2.5%) 3 (5.8%) NS Bentall s operation 3 (7.5%) 1 (1.9%) NS NS not significant. temperature maintained at 25 C. Cold blood cardioplegic solution was perfused continuously in a retrograde direction using a Retro-TH catheter placed in the coronary sinus, to protect the myocardium until completion of aortic arch reconstruction. Distal anastomosis of a prosthetic graft with the descending aorta or the aortic arch was done, and the arch vessels were reconstructed. Air was removed from the graft, which then was occluded. Perfusion through the graft to the distal aorta was reinstituted to resume cerebral circulation, and rewarming was started. During this time, the proximal aorta was anastomosed with the graft. After venting the prosthetic graft, the graft occlusion was released and cardiac resuscitation was begun by conventional extracorporeal perfusion through the graft or the ascending aortic return alone. The femoral arterial catheter was clamped to eliminate backflushing of emboli. Measurement of Oxygen Saturation In the frontal region of the brain, rso 2 was measured continuously throughout the operation by near-infrared spectroscopy (Invos 3100; Somanetics, Troy, MI), for which a probe was placed on the left side of the patient s forehead. By this method regional cerebrovascular oxygen saturation, a quantitive measure of hemoglobin saturation in the combined arterial, venous, and microcirculatory compartments of the brain, could be measured noninvasively. Clinical characteristics of the patients are described in Table 1. Statistical Analyses Continuous variables are presented as means ( standard deviation). Discrete variables are presented as counts and percentages. Univariate comparisons used two-tailed t tests for continuous variables and Pearson s 2 statistic for categoric variables. One-way analysis of variance was used to assess data regarding brain oxygen saturation during cerebral perfusion. Differences were considered statistically significant when p was less than Results No differences were evident in cardiopulmonary bypass time, myocardial protection time, or minimum tympanic membrane temperature, but cerebral perfusion time was minutes for RCP and minutes for SCP, and was significantly shorter for the RCP group because of technical differences in the operative approaches (Table 2). Immediately before the start of cardiopulmonary bypass (baseline value), rso 2 was 63.9% 6.9% for RCP and 66.1% 5.3% for SCP (not significantly different). As patients were core-cooled to 20 C, rso 2 gradually increased, reaching 73.1% 8.8% and 74.1% 7.9% in the RCP and SCP groups, respectively. With initiation of circulatory arrest, rso 2 decreased in a steep curve. Although rso 2 recovered to the prearrest value in the SCP group soon after the start of cerebral perfusion, the RCP group showed a nearly constant, sustained decrease in rso 2, although the rate of decrease slowed when cerebral perfusion was initiated. At the end of brain protection, rso 2 was 57.4% 12.2% (lowest value, 36%) for RCP and 71.7% 6.9% (lowest value, 63%) for SCP (Fig 1). Decreases during the interval from initiation to conclusion of brain protection were 24.4% and 3.7%, respectively, for RCP and SCP groups. These data indicated that brain tissue oxygenation deteriorated significantly in the RCP group; rso 2 decreased to the baseline value about 25 minutes after the start of brain protection, and at the end of brain protection rso 2 averaged 0.89 of the baseline value in contrast to 1.08 for SCP. No difference in changes of rso 2 existed between the different RCP methods (hypothermic circulatory arrest versus femoral artery perfusion; Fig 2). Clinically, three patients in the RCP group had not Table 2. Comparison in Operation Data Between Retrograde Perfusion and Selective Perfusion Operative Variable perfusion time (min) Cardioplegic time (min) Cardiopulmonary bypass time (min) Minimum tympanic temp ( C) Minimum rectal temp ( C) Retrograde Perfusion a (n 40) a Values are mean standard deviation. NS not significant. Selective Perfusion a (n 52) p Value NS

4 1094 HIGAMI ET AL Ann Thorac Surg RETROGRADE PERFUSION AND BRAIN OXYGEN 1999;67: Fig 1. Changes in regional cerebrovascular oxygen saturation measured by near-infrared optical spectroscopy during retrograde cerebral perfusion or selective cerebral perfusion. awakened by the next day despite absence of focal neurologic deficits. Their rso 2 values at the end of brain protection were 36%, 41%, and 48%, and their cerebral protection times were 82, 91, and 55 minutes. Their respective ratios of rso 2 to baseline value were 0.67, 0.64, and 0.63, all less than 0.7. The second of these patients was found to have cerebral infarction, most likely caused by hypoperfusion to the brain. These 3 unresponsive patients were among 5 patients whose ratios of rso 2 at the end of brain protection to baseline value were 0.7 or less. Three patients in the SCP group had small cerebral emboli but survived. Their rso 2 values at the end of brain protection were 79%, 71%, and 63%, and their cerebral protection times were 144, 136, and 190 minutes. Because their rso 2 ratios were 1.22, 1.20, and 1.19, cerebral infarction in these three SCP patients was considered to be caused not by cerebral ischemia, but most likely by embolism of debris. Fig 2. Changes in regional cerebrovascular oxygen saturation ratio (rso 2 ratio [postoperative rso 2 value to baseline value]) with retrograde cerebral perfusion. Baseline value means the rso 2 value before cardiopulmonary bypass. Fig 3. Regional cerebrovascular oxygen saturation (rso 2 )inapatient who had 54 minutes of retrograde cerebral perfusion after 13 minutes of circulatory arrest at a tympanic membrane temperature below 20 C. This graph illustrates typical changes in rso 2 during retrograde cerebral perfusion. The arrow X indicates when temporary circulatory arrest began. The arrow Y indicates when retrograde cerebral perfusion started. The arrow Z indicates the start of reperfusion. (HCA hypothermic circulatory arrest.) Comment Profound hypothermic systemic circulatory arrest and selective cerebral perfusion have been used to protect the brain during operations involving the aortic arch. The goal of brain protection is to prevent both cerebral ischemia and embolism due to debris. Circulatory arrest in which the aorta and its branches are not clamped carries a low risk of embolism, but cerebral ischemia gradually develops even under conditions of profound hypothermia, so this method has strict time limitations. The safe duration of hypothermic circulatory arrest generally has been estimated to be 40 to 45 minutes even at a rectal temperature of less than 18 C [11 13]. Hypothermic SCP reduces basic metabolic requirements and provides an uninterrupted oxygen and nutrient supply to the brain. Therefore, SCP has less-strict time limitations, and it may prevent ischemic damage to the brain even in patients requiring relatively lengthy reparative procedures [14, 15]. However, this technique requires complicated procedures to obtain arterial access for extracorporeal circulation and cumbersome dissection of the aortic arch vessels. Selective cerebral perfusion also presents problems such as embolism of debris resulting from clamping as well as other complications. We considerably reduced the frequency of cerebral complications by using nonclamping selective cerebral perfusion [4]. Magnetic resonance spectroscopy performed in pigs by Filgueiras and coworkers [2] showed that the small amount of blood flow provided by retrograde cerebral perfusion during hypothermic circulatory arrest was unable to prevent metabolic evidence of ischemia, such as a significant decrease of adenosine triphosphate and creatine phosphate accompanied by accumulation of inorganic phosphate. An experimental study by Ye and coworkers [3] showed that retrograde perfusion during deep hypothermic circulatory arrest provided better ce-

5 Ann Thorac Surg HIGAMI ET AL 1999;67: RETROGRADE PERFUSION AND BRAIN OXYGEN 1095 rebral protection than deep hypothermic circulatory arrest alone, but nonetheless offered less brain protection than antegrade cerebral perfusion during prolonged deep hypothermic circulatory arrest (120 minutes). Time limitations of circulatory arrest can be relaxed somewhat by combining arrest with RCP. Nonetheless a safe time limit of 60 to 80 minutes has been proposed [5 8, 16]. In another report, safe duration of RCP varied between individuals, and cerebral complications occurred in 21% of patients, even when the duration of perfusion was kept within 80 minutes [8]. No unanimity of opinion exists as to the safe time limit for cerebral perfusion. In the present study we compared the safe time range of cerebral perfusion between RCP and SCP from the standpoint of brain tissue oxygenation by measuring regional cerebrovascular oxygen saturation with an Invos 3100 optical spectroscope. For brain protection we chose one of three methods including RCP through the superior vena cava with hypothermic circulatory arrest [5], RCP through the femoral artery with the patient in the Trendelenburg position [6], and nonclamping SCP [4]. Our choice considered both the extent of aortic arch reconstruction required and the surgical approach (median sternotomy or lateral thoracotomy). We preferred nonclamping SCP for total arch replacement with individual reconstruction of the arch vessels via a median sternotomy, RCP with femoral artery perfusion for distal arch replacement via a left lateral thoracotomy, and RCP with hypothermic circulatory arrest for proximal arch or ascending aorta replacement, especially in aortic dissection. Near-infrared light at a wavelength of 650 to 1100 nm is not easily scattered by visible light and is attenuated only by particular biologic pigments, such as oxyhemoglobin, deoxyhemoglobin, and cytochrome oxidase [17, 18]; few other pigments exist in large quantities in the brain. The optical spectroscope determines the concentration of pigments in the brain tissue from their near-infrared absorbance [9]. The Invos 3100 optical spectroscope can assess only a limited region and cannot monitor the entire brain. This limitation of the method is a subject for debate [10]. Advantages of the method for intraoperative monitoring include noninvasiveness and real-time operation. Whereas the rate of decrease in rso 2 during SCP determined by optical spectroscopy was 0.02% per minute, it was significantly greater (0.41% per minute) for RCP and continued to decrease linearly over time, indicating progression of cerebral ischemia during RCP. The time required for decrease to the baseline rso 2 value, as converted from the mean gradient of decrease, was about 25 minutes. A question exists as to how long cerebral ischemia can be kept reversible once the baseline value is reached. In the present study 3 patients in the RCP group whose rso 2 values at the end of brain protection were 36%, 41%, and 48% had delayed awakening. Their respective rso 2 ratios were 0.67, 0.64, and 0.63, all less than 0.7. Two of the 3 patients had cerebral perfusion for over 70 minutes. In contrast, the 2 other patients who had RCP for 70 minutes or more awakened without delay but had rso 2 ratios of 0.78 and 0.89, both over 0.7. Therefore, cerebral perfusion procedures exceeding 70 minutes might induce ischemic brain damage, but this is not invariable provided that the rso 2 ratio remains higher than 0.7. Conversely, 5 patients, all in the RCP group, ended the period of brain protection with rso 2 ratios of 0.7 or less. Two of these patients, both with ratios of 0.70 but short RCP times (24 and 35 minutes), had no neurologic deficits. The other 3 patients had delayed awakening as mentioned previously (Fig 2). The frequency of ischemic brain damage in patients with an rso 2 ratio less than 0.7 was considerable (60%). Consequently, rso 2, measured noninvasively in real time, is a useful warning that ischemic brain damage during cerebral protection is likely if the procedure is prolonged. McCormick and associates [19] investigated the relationship between rso 2 and electroencephalographic changes after a 7% oxygen load by using near-infrared spectroscopy and found that slow waves appeared at normothermia when rso 2 was less than 55%. Ausman and associates [20] operated on patients with cerebral aneurysms using circulatory arrest at 18 C. They found that 5 patients whose rso 2 remained above 35% had no neurologic injury attributable to hypoxia, but 2 patients with a lower rso 2 (30% and 34%) failed to regain consciousness postoperatively. Postmortem examination of the 2 patients showed cellular changes consistent with diffuse cerebral ischemia. No consensus has been reached as to the minimum rso 2 required to prevent irreversible cerebral injury. Assuming the safe limit of rso 2 to be 45% in a hypothermic environment (tympanic membrane temperature 18 to 20 C), a time limit of about 70 minutes can be calculated from the average rate of decrease in rso 2 found in the present study. This time limit concurs with the safe time range proposed by many investigators. If a reliable minimum tolerable limit for rso 2 can be established by this method, it should be possible to predict safe time limits for individual patients of rso 2 on the basis of the spectroscopically determined gradient of decrease in rso 2, to prevent the development of excessive cerebral ischemia under conditions of profound hypothermia. In conclusion, SCP has no limitation within the usual times used. With RCP, on the other hand, rso 2 continues to decline with time decreasing below the baseline value after as few as 25 minutes of brain protection. Our data indicate that an rso 2 ratio of 0.7 is a critical minimum level. References 1. Mujsce DJ, Towfighi J, Yagar JY, Vannucci RC. Neuropathologic aspects of hypothermic circulatory arrest in newborn dogs. Acta Neuropathol 1993;85: Filgueiras CL, Winsborrow B, Ye J, et al. A 31 P-magnetic resonance study of antegrade and retrograde cerebral perfusion during aortic arch surgery in pigs. J Thorac Cardiovasc Surg 1995;110: Ye J, Yang L, Del Bigio MR, et al. Neuronal damage after hypothermic circulatory arrest and retrograde cerebral perfusion in the pig. Ann Thorac Surg 1996;61:

6 1096 HIGAMI ET AL Ann Thorac Surg RETROGRADE PERFUSION AND BRAIN OXYGEN 1999;67: Ogawa K, Higami T, Asada T, et al. Aortic arch reconstruction without aortic cross-clamping using separate extracorporeal circulation. J Jpn Thorac Surg 1993;41: 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: Takamoto S, Matsuda T, Harada M, Shimamura Y, Miyata S. Simple hypothermic retrograde cerebral perfusion during aortic arch surgery. J Cardiovasc Surg 1992;33: Yoshimura N, Okada M, Ota T, Nohara H. Pharmacologic intervention for ischemic brain edema after retrograde cerebral perfusion. J Thorac Cardiovasc Surg 1995;109: Yamaki F, Hashimoto A, Aomi S, et al. Effect of retrograde cerebral perfusion in preserving cerebral function during circulatory arrest. J Jpn Thorac Surg 1993;41: McCormick PW, Stewart M, Ray P, Lewis G, Dujovny M, Ansman JI. Measurement of regional cerebrovascular haemoglobin oxygen saturation in cats using optical spectroscopy. Neurol Res 1991;13: Brown R, Wright G, Royston D. A comparison of two systems for assessing cerebral venous oxyhaemoglobin saturation during cardiopulmonary bypass in humans. Anesthesia 1993;48: Deeb GM, Jenkins E, Bolling SF, et al. Retrograde cerebral perfusion during hypothermic circulatory arrest reduces neurologic morbidity. J Thorac Cardiovasc Surg 1995;109: Bigelow WG, Callaghan JC, Hopps JA. General hypothermia for experimental intracardiac surgery. Ann Surg 1950;132: Mezrow CK, Sadeghi AM, Gandsas A, et al. blood flow and metabolism in hypothermic circulatory arrest. Ann Thorac Surg 1992;54: Kazui T, Kimura N, Komatsu S. Surgical treatment of arch using selective cerebral perfusion. Eur J Cardiothorac Surg 1995;9: Bachet J, Guilmet D, Goudot B, et al. Cold cerebroplegia. A new technique of cerebral protection during operation on the transverse aortic arch. J Thorac Cardiovasc Surg 1991; 102: Usui A, Abe T, Murase M. Early clinical result of retrograde cerebral perfusion for aortic arch operation in Japan. Ann Thorac Surg 1996;62: Delpy DT, Cope M, Zee P, Arridge S, Wray S, Wyatt J. Estimation of optical pathlength through tissue from direct time of flight measurement. Phys Med Biol 1988;33: Chance B, Leigh JS, Miyake H, et al. Comparison of timeresolved and -unresolved measurement of deoxyhemoglobin in brain. Proc Natl Acad Sci USA 1988;85: McCormick PW, Stewart M, Goetting MG, Balakrishnan G. Regional cerebrovascular oxygen saturation measured by optical spectroscopy in humans. Stroke 1991;22: Ausman JI, McCormick PW, Stewart M, et al. oxygen metabolism during hypothermic circulatory arrest in humans. J Neurosurg 1993;79: INVITED COMMENTARY The quest for a way to reliably determine the safe duration of circulatory arrest continues. The study by Higami and associates is a clinical study supporting the use of cerebral oxygen saturation measurements as one method of monitoring the status of the brain during circulatory arrest. The most important contribution this study makes is to provide clinical data corroborating previous animal studies showing the lack of efficacy of retrograde cerebral perfusion in providing oxygen to the brain. Many clinical reports of retrospective studies that lacked careful neurologic assessment have touted the benefits of retrograde perfusion. As yet, no carefully performed prospective clinical study of the benefits of retrograde perfusion has been published. Despite the laboratory and clinical studies that do not show benefit, retrograde perfusion might have a role as an adjunct for neuroprotection by flushing out potential air and particulate emboli from the arch vessels. Julie A. Swain, MD Division of Cardiothoracic and Vascular Surgery University of Kentucky College of Medicine 800 Rose St, MN 276 Lexington, KY Jswain@pop.uky.edu 1999 by The Society of Thoracic Surgeons /99/$20.00 Published by Elsevier Science Inc PII S (99)