|
|
||||||||
Ann Thorac Surg 2001;72:1774-1782
© 2001 The Society of Thoracic Surgeons
a Department of Anesthesiology, Mount Sinai School of Medicine, New York, New York, USA
b Department of Cardiothoracic Surgery, Mount Sinai School of Medicine, New York, New York, USA
* Address reprint requests to Dr Reich, Department of Anesthesiology, Box 1010, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029-6574, USA
e-mail: david.reich{at}mssm.edu
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
The use of RCP was originally reported by Mills and Ochsner for the management of massive arterial air embolism during cardiopulmonary bypass (CPB) in 1980 [1]. In 1982 Lemole and colleagues described intermittent RCP as a method of facilitating intraluminal graft placement in the aorta [2]. In 1990 Ueda and associates first described the routine use of continuous RCP in thoracic aortic surgery for the purpose of cerebral protection during the period of obligatory interruption of anterograde cerebral flow [3].
The mechanisms whereby RCP may accomplish neuroprotection include providing cerebral metabolic support, expelling atheromatous and gaseous emboli from the cerebral vasculature, and maintaining cerebral hypothermia. Conversely, there is evidence that RCP may worsen neurological outcome, most likely by inducing cerebral edema. None of the proposed mechanisms of cerebral protection or injury noted above are definitively established, based on the conflicting results of clinical and animal laboratory studies. This review is intended to summarize the existing literature regarding RCP and cerebral protection. The controversies regarding RCP will be highlighted, given the discordant findings in the literature.
| Technique of RCP |
|---|
|
|
|---|
The cardiopulmonary bypass technique usually includes bicaval venous cannulation, with the arterial line containing Y-connectors with limbs to the venous line that are clamped during anterograde cerebral flow. Typically, at the time of institution of RCP, the superior vena cava cannula is snared, anterograde flow is interrupted, the arterial cannula is clamped, and the limb connecting the arterial return line to the superior vena caval cannula is opened. Perfused blood is returned to the oxygenator via cardiotomy suction placed in the open thoracic aorta, the surgical field, and via the inferior vena cava cannula. In some institutions, the retrograde perfusion technique may include the entire venous system using a single atrial cannula. Retrograde cerebral perfusion may be applied continuously during interruption of anterograde cerebral flow or intermittently. Some clinicians apply RCP very briefly at the end of an HCA period solely for the purpose of flushing embolic debris from the cerebral vasculature.
Retrograde cerebral perfusion is generally carried out with monitoring of the venous pressure via either a central venous catheter in the superior vena cava or a jugular bulb catheter. In most centers, flow is adjusted to maintain pressure in the range of 15 to 25 mm Hg, but venous perfusion pressures up to 40 mm Hg have been reported. For the purpose of monitoring the efficacy of cerebral perfusion during RCP, near-infrared spectroscopy, transcranial Doppler assessment of middle cerebral artery, and central retinal artery flow have been used [7, 8].
Upon reinstitution of anterograde cerebral perfusion and completion of the aortic repair, gradual warming is carried out by means of CPB. The gradient between blood and body core temperature is usually limited to less than 10°C, with a maximum blood temperature of 37°C. A warming blanket is commonly used.
| Assessing efficacy of RCP |
|---|
|
|
|---|
There are various issues that confound the interpretation of human investigations of RCP. Humans may have valves in the internal jugular veins unilaterally, bilaterally, or not at all, and these venous valves may or may not be competent [13, 14]. In an anatomical human cadaver study, RCP was described as feasible via the azygous system, despite the presence of intact venous valves in the jugular veins (caval system) [15]. The limited percentage of perfused blood that returns via the brachiocephalic vessels during RCP implies that the blood is shunted away from the brachiocephalic circulation via venous collateral channels or sequestered in tissues as edema fluid.
Laboratory investigations of cerebral blood flow
Interspecies differences in cerebral venous anatomy are among the major issues confounding the interpretation of laboratory investigations of RCP. Canine models are limited by the presence of venous valves in the head and neck. Studies of RCP that use this model therefore employ bilateral cannulation of the maxillary veins, perfusion via the superior sagittal sinus, or superior vena cava cannulation with venous valve obliteration [16, 17]. Superior vena caval cannulation is possible in the pig because there are no venous valves, but pigs differ from humans in the anatomy of the cerebral venous circulation and in the proportion of flow going to the brain versus extracranial structures [18]. These interspecies differences and the widely varying experimental conditions complicate the interpretation of the following studies.
Laser Doppler flowmetry studies have shown that regional cerebral blood flow during hypothermic RCP is 2% to 20% that of preoperative levels. Sakurada and associates [19] found parietal lobe blood flow to be 2.2% of baseline in dogs perfused via the maxillary veins with the vena cavae occluded, whereas Safi and associates [20] found blood flow to be 20% of baseline in pigs perfused via the superior vena cava. Ye and associates [21] used magnetic resonance perfusion imaging to monitor flow distribution in the brains of pigs undergoing hypothermic anterograde versus retrograde perfusion. This study found that RCP provided little or no cerebral perfusion relative to SCP, and that it resulted in poor brain perfusion during the reperfusion phase.
Studies in which the proportion of perfusate returned via the aorta was measured indicate that most of the perfusate is shunted through the venous system. In baboons perfused via the internal jugular vein, aortic arch return was less than 1%, whereas 90% of the perfusate was shunted to the inferior vena cava [22]. In mongrel dogs perfused under normothermic and hypothermic conditions via the superior vena cava at an external jugular venous pressure of 25 mm Hg, 20% of the perfusate returned through the aorta and the rest drained through the inferior vena cava [23]. In a series of investigations performed at the authors institution, pigs underwent RCP via the superior vena cava or the innominate vein with and without inferior vena cava occlusion. With superior vena cava/innominate vein perfusion, 3% to 4% of the retrograde flow returned to the aortic arch in the absence of inferior vena cava occlusion [24, 25]. With inferior vena cava occlusion, the amount of aortic arch return increased significantly in one study [26] but was not significantly different in another [27].
Studies comparing regional cerebral blood flow during RCP and CPB using the hydrogen clearance method have shown that RCP provides 20% to 50% of the flow provided by CPB. In mongrel dogs, RCP delivered via the superior vena cava at 25 mm Hg provided one-half the cerebral blood flow of CPB under both normothermic and hypothermic conditions, despite the presence of intact venous valves. Two similar studies conducted by Mori [28] and Usui and associates [29] in mongrel dogs found that hypothermic RCP delivered via the maxillary veins at pressures of 20 to 25 mm Hg resulted in parietal lobe blood flows that were 50% of the hypothermic CPB level. Oohara and associates [30] studied regional cerebral blood flow in mongrel dogs during normothermic RCP at pressures of 15 to 35 mm Hg. A perfusion pressure of 25 mm Hg was optimal and resulted in blood flow that was 33% that of normothermic CPB. Nojima and associates [31] found that hypothermic RCP delivered at pressures of 10, 20, and 30 mm Hg via the maxillary veins in mongrel dogs produced parietal lobe blood flows that were 26%, 48%, and 58% of hypothermic CPB values, respectively.
One study used the hydrogen clearance method to measure subcortical parietal flow and laser Doppler flowmetry to measure cortical parietal flow [32]. Six mongrel dogs underwent CPB at 28°C, SCP at 20°C, and RCP at 20°C. RCP was performed through bilateral internal maxillary veins at a pressure of 40 mm Hg. Compared with CPB, average flow in the subcortical tissue was 56% for SCP and 44% for RCP, both statistically significantly reduced. The authors attributed the reduced flow during SCP to hypothermia. Compared with CPB, average blood flow in the cerebral cortex was 83% for SCP and 13% for RCP (significantly less than both CPB and SCP). This study may indicate a relative maldistribution of cerebral blood flow during RCP that the authors attributed to veno-venous collateral flow.
Cerebral blood flow studies have also employed the colored microsphere method, which is probably the gold standard for determining capillary flow. Oohara and associates [30] found that cerebral flow estimates were similar to those obtained using the hydrogen clearance method (33% of CPB flow) using 50-µm microspheres, normothermia, and maxillary vein perfusion at 25 mm Hg. However, Usui and associates [33] reported regional cerebral blood flow of 13.3% to 24.5% with 50-µm microspheres, and 1.6% to 3.7% with 15-µm microspheres, compared with normothermic CPB. Estimates of cerebral blood flow are probably most accurate with 15-µm microspheres, as they are trapped in capillaries only, whereas larger spheres are trapped in collateral venous vessels that are not the site of gas exchange that supports cellular respiration.
In a study of RCP via concomitant superior vena cava and superior sagittal sinus perfusion, the authors reported that RCP provides optimal flow with driving pressures of 25 to 35 mm Hg, based on measures of brain tissue blood flow (colored microspheres), regional brain pH (photometry), and vertebral surface blood flow [34].
Two other studies also addressed capillary flow during RCP. Katz and associates [35] found that technetium 99N labeled macroaggregated albumin resulted in very poor tracer trapping under normothermic conditions and no tracer trapping under hypothermic conditions in rabbits. Ye and associates [36] compared capillary staining by an India ink solution during SCP and RCP; with RCP, only 10% of the capillaries were filled.
The large body of animal literature described above may be summarized as follows. Overall intracranial flow is probably in the range of 20% to 60% of values achieved during hypothermic CPB. While higher flows may be achievable, the potential for inducing cerebral edema limits the perfusion pressure that may be applied. The goals of maintaining cerebral hypothermia and washout of embolic debris are plausible given these overall intracranial flow data. Based on the capillary flow data, however, it is doubtful that RCP provides perfusion that supports cerebral metabolism.
Human investigations of cerebral metabolism
Several investigators have measured the oxygen extraction ratio (OER) during RCP. In a case series Cheung and associates [37] found that the OER increased at the onset of RCP in all patients, but this effect was less pronounced in patients with previous strokes and intraoperative strokes. Ganzel and associates [8] compared HCA and RCP in a small sample of patients, finding that regional cerebral oximetry was comparable between groups. However, Higami and associates [38] found that regional cerebral oximetry tended to decrease during the conduct of RCP. In a study comparing RCP in patients at two levels of hypothermia with HCA and in awake controls, more hypothermic RCP resulted in higher OER compared with awake controls [39]. The same investigators reported a case series of RCP, in which OER, pyruvate, and lactate levels during RCP were comparable to those during CPB [10]. Ueda and associates [40] were among the first to report that oxygen extraction and carbon dioxide elimination occurred during RCP. Sasaguri and associates [41] noted that OER was constant over 90 minutes of RCP. A major limitation of these studies is the difficulty in determining the extent to which the metabolic indices are related to cerebral or extracranial tissues.
Laboratory investigations of cerebral metabolism
Filgueiras and associates performed two studies of high-energy phosphates and intracellular pH using magnetic resonance in pigs with RCP at 15°C [42] and 28°C [43], respectively. In both, the pH declined during RCP compared with control groups, but it remained significantly lower during reperfusion in the second study only. In the first study, high-energy phosphate levels declined during RCP and recovered with resumption of anterograde flow (similar to three of the HCA animals). In the second study, adenosine triphosphate levels decreased during RCP and remained low during recovery. Conversely, in canine models, Mori [28] and Nojima and associates [31] found higher ATP levels during RCP than during HCA using spectrophotometry. This issue remains unresolved because of conflicting data from equally valid studies.
Several investigations have compared cerebral oxygen consumption during RCP with that measured during CPB. Oohara and associates [30] found that oxygen consumption reached a plateau at 25 mm Hg of normothermic RCP that was about one-third the level measured during normothermic CPB. Juvonen and associates [27] reported cerebral oxygen consumption of 15% during hypothermic RCP with inferior vena caval occlusion and 21% during hypothermic RCP without inferior vena caval occlusion. They obtained similar results in a subsequent study [26]. In two studies noted earlier, Usui and associates [23, 29] calculated cerebral oxygen consumption of 25% to 33% of the level achieved with hypothermic CPB. Midulla and associates [24] reported cerebral oxygen consumption during hypothermic RCP to be 21% of that compared during hypothermic SCP. To summarize these studies, cerebral oxygen consumption during RCP appears to lie in the range of 15% to 33% of that provided by anterograde flow, independent of perfusate temperature.
In a study comparing hypothermic RCP with a normothermic preoperative baseline, Sakurada [19] found that cerebral oxygen consumption was only 3.3% of the baseline value compared with 32% during hypothermic SCP. Using dual-wavelength spectroscopy, Safi and associates [20] found that RCP was associated with higher cerebral tissue oxygenation compared with HCA. Therefore, while RCP appears to provide inadequate metabolic support relative to normothermic baseline conditions, it also appears to provide some metabolic support when compared with HCA.
In summary, there is little doubt that there is oxygen consumption by tissues perfused during RCP. What remains unclear is what proportion of the oxygen is consumed by cerebral tissues and whether these values are sufficient to support cellular metabolism and prevent neuronal injury.
Human investigations of biochemical markers of neuronal injury
S-100B is a glial protein that is released into the cerebrospinal fluid and bloodstream during cellular injury. It has been proposed to be a marker of global neuronal injury in cardiac surgery patients. It remains uncertain, however, whether S-100B levels correlate with measures of neurological or neuropsychological deficits in patients undergoing surgery that requires CPB [4446]. Nevertheless, there are three studies of S-100B levels in patients undergoing RCP [4749], but these investigations do not significantly add to our understanding of outcomes following RCP.
Laboratory investigations of histopathology
There are heterogeneous reports of the effects of RCP on histopathology of the brain in animal models. Two reports found improved histopathology; Imamaki and associates [50] found that 120 minutes of RCP at 15°C in mongrel dogs resulted in significantly less pathology in the CA-1 field of the hippocampus than HCA, but not RCP at 20°C. Midulla and associates [24] reported that RCP was associated with cerebral injury in 1 of 6 animals versus 9 of 12 animals undergoing HCA.
Others have found neutral effects of RCP on histopathological outcome. Crittenden and associates [51] compared sheep undergoing two hours of HCA, HCA with external cranial cooling, SCP, or RCP. They found that all four groups exhibited similar neuronal degeneration. Ye and associates [52] examined the brains of pigs that underwent 120 minutes of RCP, SCP, HCA, or anesthesia alone and found that RCP and HCA resulted in injury that was more severe than SCP or anesthesia alone. The degree and pattern of injury was similar for RCP and HCA. Retrograde cerebral perfusion and HCA also resulted in decreased levels of a protein indicator of neuronal integrity (microtubule-associated protein 2) in the hippocampus, compared with SCP and control subjects [53].
Juvonen and associates reported mixed results in two papers addressing histopathological outcome [26, 27]. In the first study, histopathological injury (mild or severe) was found in no animal after SCP, in 27% after RCP, in 47% after HCA, and in 68% after RCP with inferior vena caval occlusion [27]. In the second study, cerebral embolization was induced by injecting microspheres into the cerebral circulation, and this resulted in severe histopathological injury. Despite partial embolic washout in the group of animals that underwent RCP with inferior vena caval occlusion, RCP did not result in less neuronal injury [26]. In both studies RCP with inferior vena caval occlusion was associated with increased injury compared with RCP without inferior vena caval occlusion. That RCP with inferior vena caval occlusion resulted in greater injury may have been related to more retrograde cerebral flow resulting in greater cerebral edema and increased intracranial pressure.
The following investigations suggest impaired histopathological outcome with RCP. Boeckxstaens and associates [22] found that baboons that underwent 1 hour of hypothermic RCP had greater glial edema (status spongiosus) than baboons that underwent 1 hour of hypothermic circulatory arrest. Mohri and associates [54] compared mongrel dogs undergoing RCP, SCP, or no treatment and found that RCP produced localized infarctions (as indicated by uneven carbon black staining), destruction of the bloodbrain barrier, and cerebral edema (as indicated by greater uptake of Evans blue). These observations are limited by the absence of an HCA group. Yerlioglu and associates found that RCP delivered at a high pressure (> 40 mm Hg) exacerbated neurological injury due to microembolism [25]. In the absence of embolization, RCP also was associated with increased injury.
There are only two studies that address pharmacologic adjuncts and animal histopathology related to RCP. Wang and associates [55] studied the effects of lidocaine on neuropathology in 14 dogs subjected to 120 min of RCP at 20°C. Lidocaine reduced ischemic injury in the parietal cortex, the CA-1 field of the hippocampus, the ventral posterior lateral nucleus of the thalamus, and across all regions examined. Yoshimura and associates [56] evaluated brain edema in three groups of dogs: RCP for 120 min at 20°C (control); mannitol prior to RCP; and continuous therapy with an antivasospastic drug (1,2-bis[nicotinamido]-propane). Both cerebrovascular resistance and intracranial pressure showed a marked increase in control animals after resumption of anterograde flow, and both were significantly higher than in animals in the treatment groups. Water content of the brain tissue was also significantly higher in the control group.
Laboratory investigations of behavioral outcome
In two studies from the same laboratory, Juvonen and associates reported heterogeneous results. In the first study, behavioral outcome was better in RCP without inferior vena caval occlusion versus HCA, but RCP with inferior vena caval occlusion was worse [27]. In the second study, no method of perfusion resulted in adequate behavioral recovery following embolization, and RCP with inferior vena caval occlusion resulted in minor and transient behavioral impairment in the absence of embolization [26]. Previously, the same laboratory reported that RCP and HCA (with the head packed in ice) had similar outcomes that were significantly improved compared with HCA without packing of the head in ice [24]. Safi and associates reported similar findings [20]. Yerlioglu and associates [25] reported that complete behavioral recovery occurred with both RCP and anterograde perfusion. Cerebral embolization, however, resulted in poorer recovery, especially with RCP greater than 40 mm Hg.
The above studies do not provide a consistent picture of behavioral recovery following experimental RCP. The varied application of hypothermia, RCP cannulation technique, perfusion pressure and duration, and the effects of experimental cerebral embolization complicate comparison of these studies.
Two studies by Anttila and associates [57, 58] suggest that RCP provides experimental neuroprotection by maintaining cerebral hypothermia. In the first study, pigs were perfused on CPB at 25°C prior to instituting RCP for 45 minutes at 15°C. A comparison group underwent HCA with ice-packing. More RCP animals survived and their histopathological outcome was better than the HCA animals [57]. The second study compared pigs undergoing 15°C RCP at systemic hypothermia of 25°C, 25°C RCP at systemic hypothermia of 25°C, or HCA at 25°C for 40 minutes. In the 15°C RCP group, epidural temperatures were lower during the procedure, rates of survival were higher, behavioral outcome was better, and there was less neuronal injury by histopathology [58].
Clinical outcome in thoracic aortic surgery
The relationship between use of RCP and mortality rates is unclear. In some studies, RCP duration was not found to be a predictor of death [41, 59, 60], whereas in others it was [61, 62]. Clinical outcome studies comparing RCP and HCA have also yielded mixed results. Retrograde cerebral perfusion has been found to be associated with comparable mortality rates by some investigators [43, 6264] and reduced mortality rates by others [6567]. In studies that included SCP patients, RCP patients had similar mortality rates in three studies [6870].
Studies examining postoperative central nervous system function following thoracic aortic surgery have typically assessed the incidence of neurological morbidity. A negative outcome is generally defined as one or more of the following: coma, delirium, obtundation, sensory deficit, motor deficit, and speech disturbance. The mechanism of postoperative neurological morbidity is often assumed to be embolic stroke, and some studies have included [60] or required [65, 71] in their criteria radiographic evidence of infarction. Cerebral hypoperfusion and microembolic infarction, however, are also believed to play a role.
The literature examining the relationship between RCP and neurological morbidity is mixed. In some case series, RCP duration was not found to be a predictor of neurological morbidity [40, 59, 60, 72], whereas in others it was [23, 33, 47, 62]. Retrograde cerebral perfusion has been found to be associated with neurological morbidity rates that are either similar to those associated with HCA [39, 47, 64] or lower [63, 6567, 71], especially in older patients [72]. In studies that included SCP patients, RCP patients had similar outcomes in three studies [6870], and a worse outcome in one study [71]. Several studies have reported mortality and neurological morbidity rates without reference to a comparison group [7375].
At the authors institution, a gross assessment of temporary neurological dysfunction is routinely performed in patients surviving surgery for at least 24 hours. In 91 patients with 40 to 80 minutes of "total cerebral protection time," there was less temporary neurological dysfunction in patients having SCP than in those having either HCA alone or HCA with RCP (odds ratio 0.3, p < 0.048). The incidence of temporary neurological dysfunction was 69.2% in RCP patients and 52.9% in patients having HCA alone, but this finding was not statistically significant [76].
Most studies examining postoperative neurological dysfunction only detect gross motor deficits, altered level of consciousness, and disorientation. It is well established that the incidence of neurological morbidity following CPB is far less than the incidence of neuropsychological morbidity [77]. The neuropsychological examination involves evaluation of cognitive functions such as attention, processing speed, language, visuospatial functions, memory, and executive functions. We have recently compared neuropsychological outcome in 12 RCP patients and 44 HCA patients. Relative to HCA, RCP was associated with memory dysfunction and an overall increase in neuropsychological dysfunction in multiple domains, even when controlling separately for age and cerebral ischemia time [78]. The only other study examining postoperative cognitive function in RCP patients compared Full Scale, Verbal Scale, and Performance Scale Intelligence Quotients (IQs) in 15 patients who received RCP during thoracic aortic surgery to age-matched control subjects 3 months postoperatively [39]. These authors found no difference between the groups. It is possible that the failure to detect a difference in that study was related to the limitations of the IQ tests, which do not reflect various cognitive processes [79].
A summary of the morbidity and mortality data described in the paragraphs above is presented in Table 1. A potentially confounding factor in most papers that show benefit of RCP in reducing the risk of death and neurological morbidity is that the multivariate analyses did not include year of operation in the model. Therefore, it is impossible to separate the effect of the learning curve from the effect of RCP that was usually introduced as surgeons became more proficient in the performance of the procedures. The lack of randomized, prospective, controlled clinical investigations of RCP is a major limitation. Such studies may never be performed, however, because of the complexity of the constantly evolving surgical techniques, confounding patient factors, and ethical concerns.
|
| Comment |
|---|
|
|
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. Gega, J. A. Rizzo, M. H. Johnson, M. Tranquilli, E. A. Farkas, and J. A. Elefteriades Straight Deep Hypothermic Arrest: Experience in 394 Patients Supports Its Effectiveness as a Sole Means of Brain Preservation Ann. Thorac. Surg., September 1, 2007; 84(3): 759 - 767. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. D. Touati, P. Marticho, M. Farag, D. Carmi, C. Szymanski, M. Barry, F. Trojette, and T. Caus Totally normothermic aortic arch replacement without circulatory arrest Eur. J. Cardiothorac. Surg., August 1, 2007; 32(2): 263 - 268. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Hagl, N. Khaladj, S. Peterss, K. Hoeffler, M. Winterhalter, M. Karck, and A. Haverich 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., July 1, 2004; 26(1): 73 - 80. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. T. Strauch, D. Spielvogel, A. Lauten, S. L. Lansman, K. McMurtry, C. A. Bodian, and R. B. Griepp Axillary artery cannulation: routine use in ascending aorta and aortic arch replacement Ann. Thorac. Surg., July 1, 2004; 78(1): 103 - 108. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. L. Estrera, Z. Garami, C. C. Miller III, R. Sheinbaum, T. T. T. Huynh, E. E. Porat, A. Winnerkvist, and H. J. Safi Determination of cerebral blood flow dynamics during retrograde cerebral perfusion using power M-mode transcranial Doppler Ann. Thorac. Surg., September 1, 2003; 76(3): 704 - 710. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Hagl, N. Khaladj, M. Karck, K. Kallenbach, R. Leyh, M. Winterhalter, and A. Haverich Hypothermic circulatory arrest during ascending and aortic arch surgery: the theoretical impact of different cerebral perfusion techniques and other methods of cerebral protection Eur. J. Cardiothorac. Surg., September 1, 2003; 24(3): 371 - 378. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Nishizawa, A. Usui, M. Murase, and Y. Ueda pH-stat blood gas management provides better cerebral perfusion during deep hypothermic retrograde cerebral perfusion Interactive CardioVascular and Thoracic Surgery, December 1, 2002; 1(2): 88 - 92. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. A. L. Soong, S. Uysal, and D. L. Reich Cerebral Protection During Surgery of the Aortic Arch Seminars in Cardiothoracic and Vascular Anesthesia, November 1, 2001; 5(4): 286 - 292. [Abstract] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |