HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): R. Keith Warrian Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mutch, W. A. C. Articles by Saunders, J. K. Search for Related Content PubMed PubMed Citation Articles by Mutch, W. A. C. Articles by Saunders, J. K.

Ann Thorac Surg 1997;64:695-701
© 1997 The Society of Thoracic Surgeons

---

Original Article: Cardiovascular

Cerebral Hypoxia During Cardiopulmonary Bypass: A Magnetic Resonance Imaging Study

Departments of Anesthesia and Surgery, University of Manitoba, and the Institute for Biodiagnostics/National Research Council, Winnipeg, Canada

Accepted for publication March 24, 1997.

	Abstract

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Background. Neurocognitive deficits after open heart operations have been correlated to jugular venous oxygen desaturation on rewarming from hypothermic cardiopulmonary bypass (CPB). Using a porcine model, we looked for evidence of cerebral hypoxia by magnetic resonance imaging during CPB. Brain oxygenation was assessed by T2*-weighted imaging, based on the blood oxygenation level-dependent effect (decreased T2*-weighted signal intensity with increased tissue concentrations of deoxyhemoglobin).

Methods. Pigs were placed on normothermic CPB, then cooled to 28°C for 2 hours of hypothermic CPB, then rewarmed to baseline temperature. T2*-weighted imaging was undertaken before CPB, during normothermic CPB, at 30-minute intervals during hypothermic CPB, after rewarming, and then 15 minutes after death. Imaging was with a Bruker 7.0 Tesla, 40-cm bore magnetic resonance scanner with actively shielded gradient coils. Regions of interest from the magnetic resonance images were analyzed to identify parenchymal hypoxia and correlated with jugular venous oxygen saturation. Post-hoc fuzzy clustering analysis was used to examine spatially distributed regions of interest whose pixels fol-lowed similar time courses. Attention was paid to pixels showing decreased T2* signal intensity over time.

Results. T2* signal intensity decreased with rewarming and in five of seven experiments correlated with the decrease in jugular venous oxygen saturation. T2* imaging with fuzzy clustering analysis revealed two diffusely distributed pixel groups during CPB. One large group of pixels (50% ± 13% of total pixel count) showed increased T2* signal intensity (well-oxygenated tissue) during hypothermia, with decreased intensity on rewarming. Changes in a second group of pixels (34% ± 8% of total pixel count) showed a progressive decrease in T2* signal intensity, independent of temperature, suggestive of increased brain hypoxia during CPB.

Conclusions. Decreased T2* signal intensity in a diffuse spatial distribution indicates that a large proportion of cerebral parenchyma is hypoxic (evidenced by an increased proportion of tissue deoxyhemoglobin) during CPB in this porcine model. Neuronal damage secondary to parenchymal hypoxia may explain the postoperative neuropsychological dysfunction after cardiac operations.

	Introduction

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Cognitive dysfunction after cardiac operations correlates with jugular venous oxygen desaturation after rewarming from hypothermic cardiopulmonary bypass (CPB) [1]. Brain structures at risk during CPB and giving rise subsequently to cognitive dysfunction have not been completely identified. Recently, magnetic resonance imaging methods have been developed that are sensitive to brain oxygenation. These techniques exploit the differing magnetic properties of hemoglobin depending on its oxygenation state—oxyhemoglobin is diamagnetic whereas deoxyhemoglobin is paramagnetic. The presence of paramagnetic deoxyhemoglobin in blood generates distortions in the magnetic field surrounding capillaries and venules. Gradient echo magnetic resonance imaging is sensitive to magnetic field inhomogeneities. Local distortions in the magnetic field around capillaries give rise to reduced image signal intensity depending on the relative amount of deoxyhemoglobin. Therefore, the deoxyhemoglobin within the blood acts as an endogenous intravascular contrast agent giving rise to blood oxygenation level-dependent contrast [2]. On the basis of this effect, a field called functional magnetic resonance imaging has emerged, where brain activation is monitored through regional increases in blood oxygenation [3, 4].

The blood oxygenation level-dependent effect can also be used to document decreased tissue oxygenation [5]. Stehling and colleagues [6] examined image signal intensity changes in human brain during prolonged apnea using the blood oxygenation level-dependent effect. Rostrup and associates [7] measured the changes in human brain oxygen saturation induced by hypoxia and hyperoxia with gradient echo images. Prielmeier and colleagues [8, 9] used T2*-sensitive magnetic resonance imaging to observe transient alterations in brain oxygenation in rats during brief periods of anoxia or hypoxia, and suggested using this technique to study the earliest stages of ischemia. Further work examined the use of gradient echo magnetic resonance imaging to monitor changes in cerebral blood oxygenation after pharmacologically induced vasodilation [10]. Jezzard and associates [11] performed direct spectrophotometric validation of the oxygen saturation changes, as measured indirectly, using gradient echo magnetic resonance brain images in cats subjected to anoxia, apnea, and hypercapnia and found excellent agreement between the two techniques.

In this study, we used the blood oxygenation level-dependent effect to investigate the relationship between brain parenchymal oxygenation and jugular venous oxygen saturation (SjvO₂) before, during, and after CPB using a porcine model.

	Material and Methods

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
The experiments were approved by the Committee for Animal Experimentation at the University of Manitoba and the Animal Care Committee at the Institute for Biodiagnostics/National Research Council.

Surgical Preparation
After premedication (atropine 0.6 mg and ketamine 10 mg/kg intramuscularly), 7 pigs (22 to 28 kg) were anesthetized with 2.0% end-tidal isoflurane in oxygen. An ear vein was cannulated and 0.9% NaCl administered at 10 mL • kg^-1 • h^-1 during the operation. No glucose-containing solutions were infused. The operation entailed sternotomy for ascending aortic (6.5 mm Jostra) and right atrial (28F Polystan) cannulation, retrograde cannulation of the internal jugular vein (after all extracranial branches had been ligated) for cerebral venous blood sampling and intracranial pressure monitoring, femoral arterial and venous cannulation for hemodynamic and blood gas measurements, and reflection of scalp muscles for placement of the cerebral radio frequency (RF) coil. Muscle relaxation was accomplished with intermittent doses of pancuronium bromide. Temperature was recorded from nasopharyngeal and jugular venous probes. Bypass pump arterial inflow and venous outflow temperatures were also recorded. After operation, a 30-minute stabilization period ensued when animals received barbiturates intravenously using doses previously shown to be associated with an isoelectric electroencephalogram in this model [12], followed by a continuous infusion at 10 mg • kg^-1 • h^-1. Isoflurane administration was discontinued. Baseline hemodynamic and blood gas measurements were made at this time. Blood gases were measured by a Nova Biomedical Stat Profile 7. Arterial and venous oxygen saturations and hemoglobin were measured using a Radiometer Hemoximeter OSM 3 configured for porcine blood. Heparin (10,000 to 15,000 IU) was administered until the activated clotting time was more than 500 seconds. Additional heparin (1,000 IU/h as required) was administered to maintain the activated clotting time at more than 500 seconds. The bypass circuit was primed to a total volume of 1.5 L with Pentaspan, lactated Ringer's solution, mannitol (20%), and homologous blood (250 to 500 mL). After heparin administration, apulsatile CPB began with a Bio-Medicus 520C centrifugal pump, membrane oxygenator (Terumo Capiox E), arterial filtration (J & J Intersept, 40-µm mesh), and alpha-stat acid-base management. Normothermic bypass was maintained for 30 minutes when active cooling to 28°C by heat exchanger began. Pump flow was 2.7 ± 0.3 L/min. If ventricular fibrillation did not occur, the heart was arrested with 40 mEq of KCl. The aorta was not cross-clamped. Hypothermic CPB continued for 2 hours. Rewarming to baseline temperature then occurred over a 30- to 45-minute period. The temperature gradient never exceeded 8°C between arterial inflow and jugular venous outflow temperatures. When baseline temperature was reached, the temperature was kept stable for an additional 30 minutes. The heart was not defibrillated. At completion of the experiment, the animal was injected with a lethal dose of sodium thiopental and the bypass pump stopped. The time course of the experiment, and the periods of magnetic resonance imaging is shown in Figure 1. Measurements were made at baseline (pre-CPB), during initial normothermic CPB (normo-CPB) at 30-minute intervals during hypothermic CPB, and at the end of the 2-hour hypothermic period (hypo-CPB) when the nasopharyngeal temperature had returned to baseline temperature and when the nasopharyngeal temperature had stabilized at baseline values for 30 minutes (rew-CPB) and 15 minutes after death.

View larger version (23K): [in this window] [in a new window]   Fig 1. . The time course and imaging periods of the experiments. (CPB = cardiopulmonary bypass; Hypo = hypothermic; MRI = magnetic resonance imaging; Normo = normothermic; Rew = rewarming.)

Magnetic Resonance Imaging
A Bruker 7.0 Tesla, 40-cm bore magnetic resonance scanner with actively shielded gradient coils was used. A rectangular 3 cm by 6 cm surface RF coil was used to transmit/receive. As described above, at the end of the surgical preparation, a rectangular portion of the pig's scalp was removed to expose the skull. The RF coil was then placed immediately next to the skull (midway between the frontal and occipital margins). The proximity of the RF coil to the pig's brain improved the signal-to-noise ratio in the magnetic resonance images. The pig was placed supine in a cradle support and the head (along with the RF coil) was taped in position to minimize any motion effects. The pig, once positioned, was not moved throughout the entire magnetic resonance imaging procedure.

For T2* imaging, a single slice coronal gradient echo sequence with time to repeat (TR) = 100 ms, time to echo (TE) = 30 ms, 1.2-mm slice thickness, 256 x 256 matrix, 8 averages, and an 8-cm field-of-view was used. The RF coil was designed to provide high signal-to-noise ratio over a narrow region immediately above the coil. The 6-cm length allowed acquisition of signal over the entire coronal dimension of the pig brain, whereas the narrow width of 3 cm limited the imaged volume to a single slice. The high overall signal-to-noise ratio provided by this coil allowed an increase in the spatial resolution both in-plane and across the thickness of the slice. The relatively long TE, in addition to the ultra-high magnetic field of 7.0 T, increases the sensitivity to T2* effects, and hence changes in blood oxygenation.

Fuzzy Clustering Analysis
Standard region of interest (ROI) analysis usually involves preselecting a series of ROIs (usually square or circular regions of contiguous pixels) and, in each, monitoring the change in the mean signal intensity with time. Without foreknowledge of where to place each ROI, as is the case here, a more sophisticated unsupervised analysis is needed. We used an unsupervised classification analysis called fuzzy clustering [13, 14]. This method automatically selects the ROIs by grouping pixels with similar temporal patterns of signal change. The algorithm thus identifies spatially distributed ROIs whose pixels follow similar time courses. The algorithm also plots normalized T2* signal intensity (relative to pre-CPB T2* signal intensity) versus time for each ROI.

After spatial registration, a subset of pixels from each image was selected for fuzzy clustering that excluded regions outside the brain (including the pixels entirely outside the head, in the calvarium, and in the superior sagittal sinus) to focus the analysis on effects occurring in the microvasculature of the brain. The fuzzy c-means clustering algorithm [14] was used to separate the temporal data into K number of groups (or ROIs or clusters). A fuzzy index of 0.2 was used. The algorithm maximizes the intercluster distance while simultaneously minimizing the intracluster distance. In simple terms, for the case of two clusters, the algorithm attempts to identify a group (cluster) of pixels with a similar time course (ie, minimize the intracluster distance) that is very different from the other group (cluster) of pixels with a different average time course (ie, maximize the intercluster distance).

As implemented, associated with each of the K clusters is a temporal cluster centroid and a cluster membership map. The temporal cluster centroid is simply a plot of signal intensity versus time representing the weighted mean of all the pixels belonging to the particular cluster (loosely termed the average time response of the group). The weighting used for each pixel in the cluster is related to the cluster membership map. The latter is analogous to the probability that the pixel belongs to the cluster. Thus, pixels whose time course is very/not very/somewhat similar to the time course of the entire group are, respectively, assigned a high/low/intermediate membership to a particular cluster. A membership map with values ranging between zero and one is unique to fuzzy clustering techniques. Although each pixel can contain a nonzero membership to more than one cluster, the values are constrained such that the sum of the memberships of a pixel over the K clusters must equal unity. With this constraint, the fuzzy clustering algorithm generates a fuzzy partition of the data, where pixels tend to "belong strongly" to one and only one cluster.

The number of clusters used in the algorithm is selected by the user, and generally affects the clustering results. It is known that if more than the optimal number of clusters are selected, then one or more of the clusters centroids will appear similar. If too few clusters are selected, then one or more of the most similar clusters are amalgamated. Selecting more clusters than will likely be required has been used successfully, rather than determining the optimal number of clusters. The results are then viewed to determine the significant clusters. Clusters with few pixels (<2% of total pixel count) are discarded from analysis.

Statistical Analyses
Time-related changes for hemodynamics, temperature, and blood gases were evaluated by analysis of variance for repeated measures. When analysis of variance was significant, comparisons were made with the least-squares means test. Correlations between SjvO₂ and T2* signal intensity was undertaken using linear regression analysis. Bonferroni's correction was applied (p < 0.05/n; where n = number of comparisons) when multiple comparisons were made.

	Results

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
The temperature, hemodynamic, and blood gas data from the magnetic resonance experiments are shown in Table 1. During CPB, hemoglobin levels were constant. Figure 2 shows the brain images from a single experiment. Figure 3 shows the relationship between global mean T2* signal intensity and SjvO₂. A 15% ± 5% decrease in signal intensity was seen after rewarming, compared to baseline, with a further 14% ± 5% decrease after death. In contrast, SjvO₂ increased significantly with hypothermia (130% ± 20% versus baseline), and decreased with rewarming (75% ± 23% of baseline). Figure 4 shows the relationship between T2* signal intensity and SjvO₂ after institution of CPB over the course of each experiment. There were seven measurement periods in each experiment (one during normo-CPB, four at 30-minute intervals during hypo-CPB, and two during rewarming). In five of seven experiments there was a statistically significant correlation (r = 0.816 to 0.953; r = 0.754 for p 0.05 with df = 5). In the remaining two experiments the correlation coefficients were r = 0.528 and 0.638.

View larger version (46K): [in this window] [in a new window]   Fig 2. . T2* gradient echo magnetic resonance images of porcine brain. Images were obtained at baseline (Pre-CPB), during normothermic cardiopulmonary bypass (Normo-CPB), at the end of the 2-hour hypothermic period (Hypo-CPB), at the second rewarm cardiopulmonary bypass period (Rew-CPB), and 15 minutes after death. These images were used to measure global region of interest T2* signal intensity. Note the decreased T2* signal intensity with rewarming and especially after death, a consequence of the increased fraction of paramagnetic deoxyhemoglobin. The decreased signal intensity is especially evident over the superior sagittal sinus (top dead center).

View larger version (20K): [in this window] [in a new window]   Fig 3. . Global region of interest T2* signal intensity over time (mean ± standard deviation; n = 7) versus the five time periods described in Figure 1. The corresponding jugular venous oxygen saturation (SjvO2) is shown for the same time periods except at death when blood samples were unable to be obtained. There is a significant global decline in T2* signal intensity with rewarming (15% decline from baseline) with a further 14% decline on death. The jugular venous oxygen saturation significantly increases during hypothermic cardiopulmonary bypass (130% of baseline) with a significant decrease after rewarming (75% of baseline). (Hypo-CPB = at the end of the 2-hour hypothermic period; Normo-CPB= during normothermic cardiopulmonary bypass; Pre-CPB = baseline; Rew-CPB = at the second rewarm cardiopulmonary bypass period.)

View larger version (17K): [in this window] [in a new window]   Fig 4. . Cerebral T2* signal intensity in arbitrary units (107) versus jugular venous oxygen saturation. In each experiment seven measurements of T2* signal intensity were obtained with concurrent jugular venous blood sampling during the period of cardiopulmonary bypass. In five of seven experiments there was a significant correlation between the two variables.

View larger version (94K): [in this window] [in a new window]   Fig 5. . (A) Temperature-dependent group and temperature-independent group mean normalized T2* signal intensity versus time (mean ± standard deviation; n = 6) as generated by fuzzy clustering. The identified global region of interest for each experiment was normalized in the temporal domain to allow reliable identification of temporal patterns (see text for further details). Two large groups were identified: (1) a temperature-dependent group (50% ± 13% of total pixel count per experiment) that followed the changes seen in jugular venous blood sampling over time, that is, significantly increased T2* signal intensity during hypothermia and a subsequent significant decrement, and (2) a temperature-independent group (34% ± 8% of the total pixel count) that showed a progressive significant decrease in T2* signal intensity over time. (Hypo-CPB = at the end of the 2-hour hypothermic period; Normo-CPB= during normothermic cardiopulmonary bypass; Pre-CPB = baseline; Rew-CPB = at the second rewarm cardiopulmonary bypass period.) (B) A representative example showing pixel positions for the temperature-dependent group from one experiment. (C) A representative example showing pixel positions for the temperature-independent group from the same experiment as Figure 5B. This group shows areas of the brain at risk of progressive hypoxia during cardiopulmonary bypass.

	Comment

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
In this porcine model of CPB, we used T2* imaging techniques to measure cerebral parenchymal oxygenation. Our findings indicate that a portion of the brain parenchyma is hypoxic (as evidenced by increased proportions of tissue deoxyhemoglobin) during the conduct of CPB. If the single coronal slice seen in the magnetic resonance images is representative of whole brain, then up to 35% of the cerebrum becomes progressively hypoxic during CPB. Although an indirect measure of tissue oxygenation, T2* signal intensity correlates highly with direct spectrophotometric determination of tissue oxygenation as shown by Jezzard and colleagues [11]. Using near infrared spectroscopy, Baris and associates [15] monitored regional cerebral oxygen saturation during CPB in humans. They demonstrated a 70% incidence (7 of 10 patients) where regional cerebral oxygen saturation was less than 50%, during the late stages of rewarming, for an average duration of 9.3 minutes.

T2* signal intensity can rise with increasing blood flow [16]. Rewarming increases CBF and could increase regional T2* signal intensity. However, the opposite effect, a marked decrease in signal intensity was seen with rewarming. This suggests that the decreased signal intensity seen with increased concentrations of paramagnetic deoxyhemoblobin is the predominant effect seen with rewarming. In addition, studies by Rostrup and colleagues [7] indicates that flow minimally affects T2* signal strength because (1) flow changes did not abolish the signal changes caused by changes in blood oxygenation during hyperoxia and hypoxia in humans; (2) the T2* signal intensity was closely related to the changes in arterial oxygen saturation during hypoxia; and (3) the signal changes were little affected by changes in the flip angle from 40° to 10° during signal acquisition. Similar ROI signal intensity at reduced flip angles seems to imply that inflow effects have only a minor influence on the changes in signal intensity during hypoxia.

The ROI studied focused primarily on the hemoglobin saturation of the brain microvasculature, by excluding the superior sagittal sinus and major surface vessels from analysis, therefore large flow effects should have been minimized. Increases in microvascular flow, secondary to tissue hypoxia, may have been minimal in this situation, as well. With a fixed cardiac output (during CPB) the physiologic response of increased cardiac output to hypoxia is not possible. An augmented regional blood flow in response to tissue hypoxia may have been prevented, in part, by deposition of microemboli or secondary to capillary closure [17–19].

In five of seven experiments there was a positive correlation between T2* signal intensity and SjvO₂. Our findings are in agreement with those of McDaniel and associates [20] who demonstrated a similar correlation between T2 signal intensity and SjvO₂ in infant swine. Thus, measurement of SjvO₂ after rewarming from hypothermic CPB gives a relatively global index of cerebral oxygenation.

We have used fuzzy clustering analysis to allow us to examine regional changes in cerebral oxygenation. This technique has permitted a more detailed description of the spatial distribution of tissue oxygenation than is possible from study of the global ROI T2* signal intensity (Fig 3). This is most evident at time period hypo-CPB in this figure. Global T2* signal intensity is essentially unchanged from pre-CPB and normo-CPB. Examination of Figure 5A reveals that this apparently normal global T2* signal intensity (compared to pre-CPB) has two components: (1) diffusely distributed, temperature dependent, well oxygenated, parenchyma (increased T2* signal intensity in 50% ± 13% of total pixel counts) and (2) diffusely distributed, temperature independent, poorly oxygenated, parenchyma (decreased T2* signal intensity in 34% ± 8% of pixels). The corresponding increase in SjvO₂ during hypo-CPB (as seen in Fig 3) indicates that the poorly oxygenated tissue identified by fuzzy clustering must be poorly perfused; otherwise the venous oxygen saturation would be much lower than observed [21]. This explanation is compatible with either embolic or passive capillary closure of this portion of the microvasculature. As well, after rewarming to baseline temperature, partial recruitment of the poorly oxygenated parenchyma is suggested because of the significant reduction in SjvO₂ compared to that seen at the initial normothermic CPB period.

An inverse relationship between cerebral metabolic rate for oxygen and SjvO₂ has been demonstrated (because of the left shift of the oxyhemoglobin desaturation curve with decreasing temperature) [21]. The temperature-dependent effects of tissue oxygenation seen in Figures 5A and 5B are in agreement with these observations. It is important to note that the SjvO₂ at rewarming (42% ± 16%) was significantly lower than that seen during the initial normothermic CPB period (58% ± 19%) suggesting other causes for the decrease in SjvO₂ besides increased temperature after rewarming.

The fact that similar T2* signal intensity time courses were present in each of the six experiments examined, and not present in the control experiment, suggests a common mechanism for tissue hypoxia during CPB. Such brain hypoxia, if present in humans, could contribute to the postoperative neuropsychological dysfunction, which has been correlated with jugular venous oxygen desaturation after rewarming after hypothermic CPB [1]. Such microregional brain hypoxia is compatible with the observation that cerebral edema was present after CPB in all patients, within 1 hour of the end of cardiac operation, as assessed with a different magnetic resonance imaging approach [22].

The difference in T2* signal intensity after death (see Fig 5A) requires explanation. When measured at 15 minutes after death, the same T2* signal intensity would, perhaps, be expected in the two spatially distributed ROIs. The difference is compatible with a greater concentration of deoxyhemoglobin due to a larger regional cerebral blood volume in the temperature independent ROI. An increased regional cerebral blood volume with ischemia has been shown using magnetic resonance contrast techniques [23]. As well, microcirculatory sludging with venular pooling and edema has been demonstrated with apulsatile perfusion [24]. At the end of the experiment, a lethal injection of barbiturate was immediately followed by cessation of roller pump flow. The regional cerebral blood volume differences would be preserved in such circumstances because of zero blood flow. Thus, in the temperature-dependent, well-oxygenated parenchyma, the regional cerebral blood volume would be less than in the temperature-independent, poorly oxygenated parenchyma. By 15 minutes, desaturation of hemoglobin should be of equal magnitude in both ROIs, but in greater concentration in the temperature-independent ROI at death, resulting in decreased T2* signal intensity, as is seen in Figure 5A.

With this experimental model, the microembolic load to the brain should have been minimal (nonatheromatous aorta, deep barbiturate anesthesia, alpha-stat acid base management, membrane oxygenation, arterial filtration, no aortic cross-clamping or unclamping, and no cardiac ejection on rewarming [17–19]). This experiment suggests that causes other than microembolization contribute importantly to cerebral hypoxia during CPB. Changes in cerebrovascular autoregulatory behavior after CPB, especially with rewarming, may contribute [25]. In addition, pathologic critical closure in the microvasculature, secondary to apulsatile CPB, could explain the progressive hypoxia seen. Our results could explain why, in some models, brain damage secondary to CPB, is independent of cerebral microembolic load [26, 27]. Our findings provide insights into why the brain is at risk of injury during CPB and indicate that cerebral hypoxia occurs during the conduct of CPB.

	Acknowledgments

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Doctors Mutch, Warrian, Kozlowski, and Saunders are funded by the Canadian Heart and Stroke Foundation. Doctor Mutch is funded by the Medical Research Council of Canada. We thank Dr Gyaandeo Maharajh for surgical assistance and Dr Raymond Somorjai for development work on the fuzzy clustering algorithm. We also thank Dr Ian R. Thomson for critical review of the manuscript.

	Footnotes

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Address reprint requests to Dr Mutch, Department of Anesthesia, St. Boniface General Hospital, 409 Tache Ave, Winnipeg, Manitoba, Canada, R2H 2A6 (e-mail: mutch{at}bldghsc.lan1.umanitoba.ca).

	References

Top Footnotes Abstract Introduction Material and Methods Results Comment Acknowledgments References

 

1. Croughwell ND, Newman MF, Blumenthal JA, et al. Jugular bulb saturation and cognitive dysfunction after cardiopulmonary bypass. Ann Thorac Surg 1994;58:1702–8.[Abstract]
2. Ogawa S, Lee TM. Magnetic resonance imaging of blood vessels at high fields: in vivo and in vitro measurements and image simulation. Magn Reson Med 1990;16:9–18.[Medline]
3. Ogawa S, Tank DW, Menon R, et al. Intrinsic signal changes accompanying sensory stimulation: functional brain mapping with magnetic resonance imaging. Proc Natl Acad Sci 1992;89:5951–5.[Abstract/Free Full Text]
4. Bandettini PA, Wong EC, Hinks RS, Tikofsky RS, Hyde JS. Time course EPI of human brain function during task activation. Magn Reson Med 1992;25:390–7.[Medline]
5. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci 1990;87:9868–72.[Abstract/Free Full Text]
6. Stehling MK, Schmitt F, Ladebeck R. Echo-planar MR imaging of human brain oxygenation changes. J Magn Reson Imag 1993;3:471–4.[Medline]
7. Rostrup E, Larsson HBW, Toft PB, Garde K, Henriksen O. Signal changes in gradient echo images of human brain induced by hypo- and hyperoxia. NMR Biomed 1995;8:41–7.[Medline]
8. Prielmeier F, Merboldt K-D, Hanicke W, Frahm J. Dynamic high-resolution MR imaging of brain deoxygenation during transient anoxia in the anesthetized rat. J Cereb Blood Flow Metab 1993;13:889–94.[Medline]
9. Prielmeier F, Nagatomo Y, Frahm J. Cerebral blood oxygenation in rat brain during hypoxic hypoxia. Quantitative MRI of effective transverse relaxation rates. Magn Reson Med 1994;31:678–81.[Medline]
10. Bruhn H, Kleinschmidt A, Boecker H, Merboldt K-D, Hanicke W, Frahm J. The effect of acetazolamide on regional cerebral blood oxygenation at rest and under stimulation as assessed by MRI. J Cereb Blood Flow Metab 1994;14:742–8.[Medline]
11. Jezzard P, Heineman F, Taylor J, et al. Comparison of EPI gradient-echo contrast changes in cat brain caused by respiratory challenges with direct simultaneous evaluation of cerebral oxygenation via a cranial window. NMR Biomed 1994;7:35–44.[Medline]
12. Maharajh GS, Pascoe EA, Halliday WC, et al. Neurologic outcome in a porcine model of descending thoracic aortic surgery: left atrial to femoral artery bypass vs clamp/repair. Stroke 1996;27:2095–101.[Abstract/Free Full Text]
13. Scarth G, McIntyre M, Wowk B, Somorjai R. Detection of novelty in functional images using fuzzy clustering [Abstract]. Soc Magn Reson 3rd Annual Meeting 1995;238.
14. Bezdek J. Pattern recognition with fuzzy objective function algorithms. New York: Plenum Press, 1981.
15. Baris RR, Israel AL, Amory DW, Benni P. Regional cerebral oxygenation during cardiopulmonary bypass. Perf 1995;10:245–8.
16. Duyn JH, Moonen CTW, van Yperen GH, de Boer RW, Luyten PR. Inflow versus deoxyhemoglobin effects in BOLD functional MRI using gradient echoes at 1.5 T. NMR Biomed 1995;7:83–8.[Medline]
17. Van der Linden J, Casimir-Ahn H. When do cerebral emboli appear during open heart operations? A transcranial Doppler study. Ann Thorac Surg 1991;51:237–41.[Abstract]
18. Blauth CI. Macroemboli and microemboli during cardiopulmonary bypass. Ann Thorac Surg 1995;59:1300–3.[Abstract/Free Full Text]
19. Padayachee TS, Parsons S, Theobold R, Gosling RG, Deverall PB. The effect of arterial filtration on reduction of gaseous microemboli in the middle cerebral artery during cardiopulmonary bypass. Ann Thorac Surg 1988;45:647–9.[Abstract]
20. McDaniel LB, Deyo DJ, Vertrees R, Kent TA, Quast MJ. Magnetic resonance imaging of brain venous hemoglobin desaturation during cardiopulmonary bypass in infant swine [Abstract]. Anesthesiol 1994;81:A693.
21. Dexter F, Hindman BJ. Theoretical analysis of cerebral venous blood hemoglobin oxygen saturation as an index of cerebral oxygenation during hypothermic cardiopulmonary bypass. Anesthesiology 1995;83:405–12.[Medline]
22. Harris DNF, Bailey SM, Smith PLC, Taylor KM, Oatridge A, Bydder GM. Brain swelling in first hour after coronary artery bypass surgery. Lancet 1993;342:586–7.[Medline]
23. Guckel F, Brix G, Rempp K, Deimling M, Rother J, Georgi M. Assessment of cerebral blood volume with dynamic susceptibility contrast enhanced gradient-echo imaging. J Comput Assist Tomogr 1994;18:344–51.[Medline]
24. Matsumoto T, Wolferth CC, Perlman MH. Effects of pulsatile and non-pulsatile perfusion upon cerebral and conjunctival microcirculation in dogs. Am Surg 1971;37:61–4.[Medline]
25. Mutch WAC, Sutton IR, Teskey JM, Cheang MS, Thomson IR. Cerebral pressure-flow relationship during cardiopulmonary bypass in the dog at normothermia and moderate hypothermia. J Cereb Blood Flow Metab 1994;14:510–8.[Medline]
26. Johnston WE, Stump DA, DeWitt DS, et al. Significance of gaseous microemboli in the cerebral circulation during cardiopulmonary bypass in dogs. Circulation 1993;88:319–329.[Free Full Text]
27. Blauth CI, Arnold JV, Schulenberg WE, McCartney AC, Taylor KM. Cerebral microembolism during cardiopulmonary bypass: retinal microvascular studies in vivo with fluorescein angiography. J Thorac Cardiovasc Surg 1988;95:668–76.[Abstract]

This article has been cited by other articles:

				K. Yoshitani, M. Kawaguchi, T. Okuno, T. Kanoda, Y. Ohnishi, M. Kuro, and M. Nishizawa Measurements of Optical Pathlength Using Phase-Resolved Spectroscopy in Patients Undergoing Cardiopulmonary Bypass Anesth. Analg., February 1, 2007; 104(2): 341 - 346. [Abstract] [Full Text] [PDF]

				C. D. Mazer, F. Briet, K. R. Blight, D. J. Stewart, M. Robb, Z. Wang, A. M. Harrington, W. Mak, X. Li, and G. M.T. Hare Increased cerebral and renal endothelial nitric oxide synthase gene expression after cardiopulmonary bypass in the rat J. Thorac. Cardiovasc. Surg., January 1, 2007; 133(1): 13 - 20. [Abstract] [Full Text] [PDF]

				C. W. Hogue Jr, C. A. Palin, and J. E. Arrowsmith Cardiopulmonary bypass management and neurologic outcomes: an evidence-based appraisal of current practices. Anesth. Analg., July 1, 2006; 103(1): 21 - 37. [Abstract] [Full Text] [PDF]

				T.-J. Zhang, J. Hang, D.-X. Wen, Y.-N. Hang, and F. E. Sieber Hippocampus bcl-2 and bax expression and neuronal apoptosis after moderate hypothermic cardiopulmonary bypass in rats. Anesth. Analg., April 1, 2006; 102(4): 1018 - 1025. [Abstract] [Full Text] [PDF]

				H. P. Grocott, H. M. Homi, and F. Puskas Cognitive Dysfunction After Cardiac Surgery: Revisiting Etiology Seminars in Cardiothoracic and Vascular Anesthesia, June 1, 2005; 9(2): 123 - 129. [Abstract] [PDF]

				H. M. Homi, H. Yang, R. D. Pearlstein, and H. P. Grocott Hemodilution During Cardiopulmonary Bypass Increases Cerebral Infarct Volume After Middle Cerebral Artery Occlusion in Rats Anesth. Analg., October 1, 2004; 99(4): 974 - 981. [Abstract] [Full Text] [PDF]

				S. Bar-Yosef, J. P. Mathew, M. F. Newman, K. P. Landolfo, H. P. Grocott, and The Neurological Outcome Research Group and C.A.R. Prevention of Cerebral Hyperthermia During Cardiac Surgery by Limiting On-Bypass Rewarming in Combination with Post-Bypass Body Surface Warming: A Feasibility Study Anesth. Analg., September 1, 2004; 99(3): 641 - 646. [Abstract] [Full Text] [PDF]

				Y. Sato, D. T. Laskowitz, E. R. Bennett, M. F. Newman, D. S. Warner, and H. P. Grocott Differential Cerebral Gene Expression During Cardiopulmonary Bypass in the Rat: Evidence for Apoptosis? Anesth. Analg., June 1, 2002; 94(6): 1389 - 1394. [Abstract] [Full Text] [PDF]

				W. A. C. Mutch, R. K. Warrian, G. M. Eschun, L. G. Girling, L. Doiron, M. S. Cheang, and G. R. Lefevre Biologically variable pulsation improves jugular venous oxygen saturation during rewarming Ann. Thorac. Surg., February 1, 2000; 69(2): 491 - 497. [Abstract] [Full Text] [PDF]

				W. A. C. Mutch, G. R. Lefevre, D. B. Thiessen, L. G. Girling, and R. K. Warrian Computer-Controlled Cardiopulmonary Bypass Increases Jugular Venous Oxygen Saturation During Rewarming Ann. Thorac. Surg., January 1, 1998; 65(1): 59 - 65. [Abstract] [Full Text] [PDF]

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): R. Keith Warrian Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mutch, W. A. C. Articles by Saunders, J. K. Search for Related Content PubMed PubMed Citation Articles by Mutch, W. A. C. Articles by Saunders, J. K.