In Deutschland hat sich in den letzten Jahren er− freulicherweise eine veränderte Wahrnehmung des Schlaganfalls als zum guten Teil vermeidbare und vor allem effektiv behandelbare Erkrankung durchgesetzt. Der Schlaganfall ist bereits heute die häufigste Ursache für eine erworbene Behin− derung und gehört zu den führenden Todesursa− chen. Bedingt durch den demografischen Wandel wird eine weitere Zunahme der Schlaganfallinzi− denz für die nächsten 2 Jahrzehnte erwartet: im 5−Jahres−Zeitraum von 2021±2025 voraussicht− lich um ein Drittel höher als von 2005±2010 [1]. Dies geht einher mit einer dramatischen Belas− tung der Sozialsysteme [1]. Die 4 evidenzbasier− ten Therapieformen des akuten Schlaganfalls umfassen die frühe Aspiringabe bei Hirninfark− Zusammenfassung !

As an archetype of human adult stem cells that can readily be harvested, enriched and expanded in vitro, mesenchymal stromal cells (MSC) have been reported to be of significance for regenerative medicine. The literature is replete with reports on their developmental potentials in pre-clinical model systems. Different preparative protocols have been shown to yield MSC-like cell cultures or even cell lines, from starting materials as diverse as bone marrow, fat tissue, fetal cord blood and peripheral blood. However, MSC are still ill-defined by physical, phenotypic and functional properties. The quality of preparations from different laboratories varies tremendously and the cell products are notoriously heterogeneous. The source and freshness of the starting material, culture media used, presence of animal sera, cytokines, cell density, number of passages upon culture, etc., all have a significant impact on the (1) cell type components and heterogeneity of the initial population, (2) differential expansion of specific subsets, with different potentials of the end products, and (3) long-term functional fate of MSC as well as other types of progenitor cells that are co-cultivated with them. Consequently, there is an urgent need for the development of reliable reagents, common guidelines and standards for MSC preparations and of precise molecular and cellular markers to define subpopulations with diverse pathways of differentiation and divergent potentials.

Spatial learning is known to depend on protein synthesis in the hippocampus. Whereas the role of the hippocampus in spatial memory is established, the biochemical and molecular mechanisms underlying this process are poorly understood. To comprehend the complex pattern of protein expression induced by spatial learning, we analyzed alterations in the rat hippocampus proteome after 7 days of spatial learning in the Morris water maze. Forty Wistar rats were randomized into two groups. Animals of group A learned to localize a hidden platform in the water maze. Animals of group B served as controls and spent exactly the same time in the water maze as animals of group A. However, no platform was used in this test and the rats could not learn to localize the target. After the last trial, hydrophilic proteins from the hippocampus were isolated. A proteome‐wide study was performed, based on two‐dimensional gel electrophoresis and mass spectrometry. Compared with non‐learning animals, 53 (70%) proteins were downregulated and 23 (30%) proteins were upregulated after 7 days in rats with spatial learning. The overall changes in protein expression, as quantified by the induction factor, ranged from −1.62 (downregulation to 62%) to 2.10 (upregulation by 110%) compared with controls (100%). Most identified proteins exhibit known functions in vesicle transport, cytoskeletal architecture, and metabolism as well as neurogenesis. These findings indicate that learning in the Morris water maze has a morphological correlate on the proteome level in the hippocampus.

Hinkelbein, J.1,2; Peterka, A.1,2; Schubert, C.1,2; Schroeck, H.2; Kuschinsky, W.2; Kalenka, A.1,2 Author Information

The key goal in the treatment of acute ischemic stroke is fast vessel recanalization. Thrombolysis with recombinant tissue plasminogen activator (rt-PA) is efficient in humans but mean time for recanalization is within hours. Ultrasound bio-effects has been shown to facilitate rt-PA mediated thrombolysis in peripheral arteries. We used an embolic stroke model in the rat. In all rats we induced an ischemic stroke by a selective occlusion of the middle cerebral artery with whole blood clots. From an entire collective of 54 rats 47 completed the protocol (n=7 died early). Four different groups (no treatment n=6; full dose rt-PA treatment only [10 mg/kg per body weight] n=14, half dose rt-PA treatment plus ultrasound n=10, and full dose rt-PA treatment plus ultrasound n=17) were investigated. We found a significant reduction of absolute as well as relative infarct volume in the full dose rt-PA plus ultrasound group (81±72 mm3; P<0.05) in comparison to untreated rats (253±159 mm3; P<0.05) as well as in comparison to rats treated with full dose rt-PA only (167±91 mm3; P<0.05). There were five intracranial bleedings giving a bleeding rate of 9.3%. In summary: ultrasound treatment in addition to rt-PA is more effective than single rt-PA treatment in reducing infarct volume and safe with regard to bleeding.

Background and objective: Acute isovolaemic haemodilution increases local and mean cerebral blood flow. It is not known whether a single haemodilution has a short-term effect only or whether it affects cerebral perfusion over a longer time period. In the present study, local and mean cerebral blood flow were determined in conscious rats after a 4, 24 and 48 h period following one-time haemodilution. Methods: Thirty-six rats were randomized to three untreated sham groups and three groups of haemodilution (4, 24 or 48 h, n = 6 for each group). Isovolaemic haemodilution with albumin 5% aimed to a target haematocrit of 0.2. Local cerebral blood flow was measured in 38 brain regions by the iodo-[14C]antipyrine method in conscious normothermic rats. Results: Isovolaemic haemodilution reduced haematocrit from 0.44 to 0.20. During the following 24 and 48 h periods, haematocrit remained low (0.22 and 0.21). Mean cerebral blood flow was similar in untreated sham groups (88 ± 12 after 4 h, 92 ± 11 after 24 h, 96 ± 10 mL 100 g−1 min−1 after 48 h). Haemodilution increased mean cerebral blood flow after 4 h (184 ± 11 mL 100 g−1 min−1), after 24 h (153 ± 13 mL 100 g−1 min−1) and 48 h (149 ± 15 mL 100 g−1 min−1) (P ≤ 0.05). Local cerebral blood flow increased in all 38 structures after 4 h haemodilution but decreased with time in six of 38 brain structures after 24 h and in 15 regions after 48 h (P ≤ 0.05). Conclusions: A single one-time haemodilution increased mean cerebral blood flow for 2 days. However, local adaptation of cerebral blood flow to a chronic low haematocrit occurred but was heterogeneous within the brain.

In April 2001, the world's first oxygen carrier for human use, Hemopure(R), a haemoglobin solution, was approved in South Africa for intravenous treatment of acute perioperative blood loss. Other AOCs for intravenous use are about to be licensed. These drugs are developed with the objective to reduce the need for blood transfusion and to limit the related risks in surgery and trauma associated with high blood loss. Oxygen transport by haemoglobin-based oxygen carriers (HBOC) that are currently undergoing clinical trial is also characterized by interactions with the nitric oxide system (e.g. vasoconstriction). Transport of physically dissolved oxygen by perfluorocarbon (PFC) emulsions is restricted to artificial ventilation with oxygen but distinguished by a higher oxygen transport capacity, facilitated oxygen diffusion and anti-inflammatory effects. Dependent on their various properties, AOC; will be useful in clinical medicine (blood loss, ischaemia, adjuvant therapy and cancer diagnosis, wound healing, transplant conservation, treatment of air embolism and caisson disease), diagnosis (contrast medium) and research (cell cultures).

Among the most important controversies in evidence-based medical science is the use of therapeutic hypothermia. Our knowledge about hypothermic brain protection is still limited and several crucial factors such as target brain temperature, start and duration of hypothermia treatment as well as various unknown others still remain to be explored. Recently, the use of therapeutic hypothermia from 32-34°C after cardiac arrest improved neurological outcome in an impressive way [1-3]. This was in contradiction to a large hypothermia trial (NABIS:H), funded by the National Institute of Health (NIH), [4] concerning brain injured patients - subjected to cooling - who failed to demonstrate any improved survival. Soon after publication of the latter study, some methodological aspects were criticised [5-7]. To improve the future accuracy of hypothermia studies, scientists are now being challenged to consider further factors which impact upon brain integrity during ischaemia and trauma [8,9]. Without doubt, among these are considerable changes in the relationship between cerebral blood flow (CBF) and metabolism. Soon after therapeutic hypothermia was used to slow down metabolic rate of ischaemic tissue - and thereby to reduce the structural damage to tissues - it become obvious that 'pH management' has an impact on outcome. Lung ventilation to normocapnia is temperature-dependent since the solubility of gases in blood is increased as temperature is lowered; the pH of arterial blood shifts in parallel to the neutral point of water towards alkalosis. If 'pH-stat' management is used during hypothermia, the analytical results of the blood-gas analysis are corrected for the actual lower body temperature thus resulting in a more hypocapnic status. The compensatory reduction of lung ventilation increases the partial pressure of carbon dioxide and thereby augments CBF. This relative hypoventilation is in contrast to 'α-stat' management where the arterial blood is not corrected for the actual body temperature. α-stat management (so-called alpha-stat by the constant alpha dissociation fraction of the imidazole moiety of histidine which is assumed to be responsible for a preserved enzyme and protein activity during hypothermia) mimics the unchanged ventilation of ectothermic vertebrates who do not control pH and temperature. After the publication of investigations demonstrating a reduced incidence of cerebral emboli following α-stat management during cardiac surgery [10,11], it is generally accepted that lung ventilation of adults during cardiosurgical hypothermia should be adjusted using α-stat management [12]. Since microemboli during cardiac surgery are primarily associated with the cannulation of large vessels to permit extracorporal circulation, the role of pH management in neurology and neurosurgery remains unclear. For hypothermic brain protection in these disciplines, until now, α-stat management was preferred due to reports of better preservation of the coupling between CBF and metabolism in animals [13] and patients [14,15]. All cited studies measured either global metabolism (by brain arterial-venous oxygen content difference [13]) or the CBF for either the entire brain or a brain hemisphere by 133Xe clearance of cerebral venous blood and argon wash-in technique, respectively. Independent variations and uncoupling of the global CBF from the global brain metabolism were demonstrated in cardiac surgical patients. Since no clinical study measured regional CBF, the decreased metabolism - together with increased global blood flow during pH-stat management - might have been misinterpreted as flow/metabolism mismatch. Therefore, the impact of the acid-base management on regional CBF and brain metabolism needs to be reconsidered. Recently, the dose-dependent effect of hypothermia in the range 32-37.5°C on CBF and metabolism (by the local cerebral glucose utilization (LCGU)) was investigated at a local level in non-ischaemic rat brains. It became evident that local CBF/metabolism coupling is well preserved during hypothermia down to at least 32°C - managed by pH-stat [16] - as well as its management by α-stat [17]. To emphasize the effect of acid-base management on CBF and brain metabolism under physiological conditions, pH-stat and α-stat data of the two latter studies were combined and compared. With the exception of an increased oxygen tension in comparison to normothermic animals, physiological variables were comparable between the groups (Table 1).Table 1: Physiological variables of the experimental groups (with permission from [16,17]).Cerebral glucose utilization (or brain metabolism) Compared to normothermic anaesthetized controls, mean cerebral glucose utilization during mild hypothermia (35°C) remained unchanged during pH-stat management but was reduced during α-stat management by 18% (Fig. 1). During pH-stat and 35°C, local cerebral glucose utilization was decreased only in two regions, but in 20 regions during α-stat (Table 2). During moderate hypothermia (32°C), both pH management reduced mean cerebral glucose utilization (α-stat by 41% and pH-stat by 50%). Local cerebral glucose utilization was reduced during pH-stat in 35 instances and during α-stat in 32 instances out of 41 brain structures. Differences between the pH management on each temperature step were a lower local cerebral glucose utilization in 13 brain structures during mild hypothermia, and an increased local cerebral glucose utilization in two brain structures (cerebellar white and in the reticular part of the substantia nigra) during mild hypothermia by α-stat treatment.Figure 1: Effects of the pH management during graded hypothermia on cerebral glucose utilization (CGU) if either α-stat (▪) or pH-stat () management is used (with permission from [16,17]).Table 2: Local cerebral glucose utilization of the experimental groups (with permission from [16,17]).Cerebral blood flow The mean CBF of anaesthetized controls was unchanged during pH-stat management but was decreased in the α-stat group by 27% during mild hypothermia (35°C) and by 44% during moderate hypothermia (Fig. 2). Local cerebral blood flow was measured in 41 different brain structures (Table 3). Compared to normothermic anaesthesia, local cerebral blood flow during mild hypothermia (35°C) increased in 13 brain structures with pH-stat management but decreased in 22 structures with α-stat management. During moderate hypothermia (32°C), local CBF decreased in one and increased in two brain structures with pH-stat, but decreased in 31 structures with α-stat (Table 2). Local CBF differences between the pH management strategies were found during mild hypothermia in all structures except for three (cochlear nuclei, hippocampus CA1, and parietal cortex). Compared to pH-stat management, moderate hypothermia using α-stat management was associated with a decrease in local CBF in all but five brain structures (cochlear nuclei, dentate nuclei, superior olive, internal capsule, medial habenulae).Figure 2: Effects of the pH management during graded hypothermia on cerebral blood flow (CBF) if either α-stat (▪) or pH-stat (▒) management is used (with permission from [16,17]).Table 3: Local cerebral blood flow of the experimental groups (with permission from [16,17]).Coupling of blood flow to metabolism The coupling of CBF to cerebral glucose utilization was maintained during mild and moderate hypothermia with α-stat management as shown by the relationship of local cerebral glucose utilization in each of the 41 brain regions to the local CBF in the same regions (Fig. 3) (P > 0.05). During pH-stat management at 35°C and 32°C, the relationship between local CBF and local cerebral glucose utilization was preserved but was shifted to a higher level (Fig. 3) (P ≤ 0.05). This means that, compared to normothermic anaesthesia, hypothermia and pH-stat management reduced the local CBF less than the local cerebral glucose utilization in a brain structure. During α-stat management, the relationship between local CBF and local cerebral glucose utilization is unchanged during hypothermia; this results in a reduction of local CBF in parallel to the local cerebral glucose utilization. In conclusion, pH management increases mean and local CBF but does not uncouple CBF from metabolism. Only the relationship of CBF to cerebral glucose utilization is shifted to a higher level during pH-stat management.Figure 3: Effects of the pH management during graded hypothermia on the coupling of cerebral glucose utilization with cerebral blood flow if either α-stat (⊕ (35°C) and ⊠ (32°C)) or pH-stat (• (35°C) and ▪ (32°C)) management is used (with permission from [16,17]). The overall relationship between local cerebral glucose utilization (LCGU) and local cerebral blood flow (LCBF) in the examined structures of the brain was assessed by the least squares fit of the data to y = ax + b, where x is the mean LCGU in a given region, and y is the mean LCBF in that same area. Normothermic anaesthetized: y = 2.4x + 37.9, r = 0.92; α-stat 35°C: y = 2.2x + 22.2, r = 0.95; α-stat 32°C: y = 2.1x + 26.4, r = 0.96; pH-stat 35°C: y = 3.4x + 31.6, r = 0.92; pH-stat 32°C: y = 3.4x + 62.4, r = 0.96. Contrasts of slopes of the LCBF/LCGU regression lines were tested by common t-test statistics with Bonferroni correction for multiple comparisons (P ≤ 0.05 between the slopes of the pH-stat groups vs. normothermia and pH-stat vs, α-stat, P > 0.05 between the slopes of the hypothermia groups at 35°C and 32°C and both α-stat groups vs. normothermia). Because of the limitations of this kind of analysis, an additional more rigorous statistical approach using log-transformed data was applied, examining the relationship of LCBF and LCGU by a repeated measure of the analysis of variance according to McCulloch et al. [23] and Ford et al. [24]. For this analysis, a computer software package (BMDP2v®; BMDP Statistical Software Inc., Los Angeles, CA, USA) considering inter-animal variability and enabling the detection of heterogeneities in the relationship between LCGU and LCBF was used.Is acid-base management relevant for the design of studies on hypothermic neuroprotection? Those results derived from experimental studies cannot be directly transferred to the clinical setting, because results were obtained from healthy animals without brain ischaemia or trauma. However, there remains the possibility that pH-stat may be superior to α-stat management for neuroprotection during moderate hypothermia: first, an abolished vessel reactivity to physiological stimuli appears only to occur in the ischaemic core whereas it wanes with distance in the perifocal area [18]. During the first hours, neuronal injury in this perifocal area is essentially reversible and may be salvageable. Therefore, preserved vascular reactions may, by vasodilatation, delay or even avoid the progressive metabolic deterioration of the penumbra and the enlargement of the dense ischaemia zone. Working on both metabolism increase and perfusion decrease, hypothermia should be better combined with an increase in CBF during pH-stat than the reduction observed during α-stat management. Another potential mechanism - in addition to better cerebral oxygenation - might be that an increased or maintained CBF itself might result in a more homogenous temperature profile throughout the brain [19] thus counteracting the considerable temperature gradient in ischaemic brain [20]. Moreover, it is well known that experimental hypothermic neuroprotection depends not only upon reduced metabolism but on several other factors such as the release of neurotoxic mediators or the integrity of the blood-brain barrier [21]. At least in modestly perfused tissues, e.g. the penumbra, one can speculate that these toxic mediators are eliminated more efficiently by a higher blood flow. On the other hand, a preserved or even an increased CBF has to be avoided so as not to reach a critically raised intracranial pressure (ICP). However, ICP elevations in brain trauma [4] and severe stroke [22] are less frequent during hypothermia than during normothermia. In conclusion, pH management does not uncouple CBF from metabolism but pH-stat management shifts the relationship of CBF to cerebral glucose utilization to a higher level. Relevant clinical outcome studies on the effect of the acid-base management during hypothermia on ischaemia or brain injury are lacking. As long as they do not exist, the pH-management used in studies for neuroprotective hypothermia needs to be carefully considered to avoid further controversial results. T. Frietsch Department of Anaesthesiology and Critical Care Medicine; Faculty of Clinical Medicine Mannheim; University of Heidelberg, Germany A. Piepgras Department of Neurosurgery; Faculty of Clinical Medicine Mannheim; University of Heidelberg, Germany P. Krafft Department of Anaesthesiology; University of Vienna, Austria S. Schwab Department of Neurology; Faculty of Clinical Medicine Heidelberg; University of Heidelberg, Germany W. Kuschinsky Department of Physiology and Pathophysiology; University of Heidelberg, Germany K. F. Waschke Department of Anaesthesiology and Critical Care Medicine; Faculty of Clinical Medicine Mannheim; University of Heidelberg, Germany

Glucose, the major fuel in the brain, is transported across the cell membranes by facilitated diffusion mediated by glucose transporter proteins. Essentially two types of glucose transporters are localized in the membranes of brain endothelial cells, astrocytes, and neurons. Their densities are well adjusted to changes in local energy demand.

We addressed the question to which extent cerebral blood flow (CBF) is maintained when, in addition to a high blood viscosity (Bvis) arterial oxygen content (CaO2) is gradually decreased. CaO2) was decreased by hemodilution to hematocrits (Hct) of 30, 22, 19, and 15% in two groups. One group received blood replacement (BR) only and served as the control. The second group received an additional high viscosity solution of polyvinylpyrrolidone (BR/PVP). Bvis was reduced in the BR group and was doubled in the BR/PVP. Despite different Bvis, CBF did not differ between BR and BR/PVP rats at Hct values of 30 and 22%, indicating a complete vascular compensation of the increased Bvis at decreased CaO2. At an Hct of 19%, local cerebral blood flow (LCBF) in some brain structures was lower in BR/PVP rats than in BR rats. At the lowest Hct of 15%, LCBF of 15 brain structures and mean CBF were reduced in BR/PVP. The resulting decrease in cerebral oxygen delivery in the BR/PVP group indicates a global loss of vascular compensation. We concluded that vasodilating mechanisms compensated for Bvis increases thereby maintaining constant cerebral oxygen delivery. Compensatory mechanisms were exhausted at a Hct of 19% and lower as indicated by the reduction of CBF and cerebral oxygen delivery.

All volatile anaesthestics depress cerebral metabolism, increase intracranial pressure and contain the potential to increase cerebral blood flow. This potential is higher for halothane anaesthesia in comparison to isoflurane-, sevoflurane- and desflurane-anaesthesia and higher during nitrous oxide- in comparison to xenon- anaesthesia. Due to coupling of cerebral blood flow to a decreased cerebral metabolism, humans exhibit mostly no change or even a decrease of cerebral blood flow during anaesthesia with contemporary volatile anaesthetics. During anaesthesia with volatile anaesthetics CO2-reactivity is preserved, within concentrations up to 1 MAC also autoregulation of cerebral blood flow.

BACKGROUND:The hypothesis of a compensatory dilation of cerebral vessels to maintain cerebral blood flow at a high blood viscosity was tested during hypercapnia in the study after replacement of blood by hemoglobin solutions of defined viscosities. If compensatory vasodilation exists at normocapnia at a high blood viscosity, vasodilatory mechanisms may be exhausted when hypercapnia is added, resulting in a lack of increase in cerebral blood flow at hypercapnia.METHODS:In conscious rats, blood was replaced by ultrapurified cross-linked hemoglobin solutions that had defined and shear rate-independent low or high viscosities (low- and high-viscosity groups). Blood viscosity differed threefold between both groups (1.2 vs. 3.6 mP x s). Thereafter, rats inhaled either a normal or an increased concentration of carbon dioxide in air. Cerebral blood flow was determined by the iodo[14C]antipyrine method.RESULTS:During normocapnia, global and local cerebral blood flows did not differ between both groups. With increasing degrees of hypercapnia, global and local cerebral blood flows were gradually elevated in the low-viscosity group (2.8 ml x mmHg(-1) CO2 x 100 g(-1) x min(-1)), whereas they remained unchanged in the high-viscosity group.CONCLUSIONS:Changes in blood viscosity do not result in changes of cerebral blood flow as long as cerebral vessels can compensate for these changes by vasodilation or vasoconstriction. However, such vascular compensatory adjustments may be exhausted in their response to further pathophysiologic conditions in blood vessels that have already been dilated or constricted as a result of changes in blood viscosity.

The present study addresses the question whether local glucose transport kinetics are correlated with local glucose transporter densities in the brain. In 47 brain structures the local rate constants for 3-O-[14C]methylglucose (3-O-MG) transport, K1 and k2, were quantified, and local glucose Glut1 and Glut3 transporter densities were determined by immuno-autoradiographic methods. Statistically significant correlations were found between the rate constants for glucose transport and the transporter densities. The correlations were tighter for Glut1 than for Glut3. Inasmuch as 3-O-MG is transported by the same transporter as glucose, these results indicate that the local densities of glucose transporters determine local glucose transport rates in the brain.

BACKGROUND:The effects of xenon inhalation on mean and local cerebral blood flow (CBF) and mean and local cerebral glucose utilization (CGU) were investigated using iodo-[14C]antipyrine and [14C]deoxyglucose autoradiography.METHODS:Rats were randomly assigned to the following groups: conscious controls (n = 12); 30% (n = 12) or 70% xenon (n = 12) for 45 min for the measurement of local CBF and CGU; or 70% xenon for 2 min (n = 6) or 5 min (n = 6) for the measurement of local CBF only.RESULTS:Compared with conscious controls, steady state inhalation of 30 or 70% xenon did not result in changes of either local or mean CBF. However, mean CBF increased by 48 and 37% after 2 and 5 min of 70% xenon short inhalation, which was entirely caused by an increased local CBF in cortical brain regions. Mean CGU determined during steady state 30 or 70% xenon inhalation remained unchanged, although local CGU decreased in 7 (30% xenon) and 18 (70% xenon) of the 40 examined brain regions. The correlation between CBF and CGU in 40 local brain structures was maintained during steady state inhalation of both 30 and 70% xenon inhalation, although at an increased slope at 70% xenon.CONCLUSION:Effects of 70% xenon inhalation on CBF in rats are time-dependent. During steady state xenon inhalation (45 min), mean values of CBF and CGU do not differ from control values, and the relation of regional CBF to CGU is maintained, although reset at a higher level.

Background Hypothermia may interfere with the relationship between cerebral blood flow (CBF) and metabolism. Because this conclusion was based on the analysis of global values, the question remains whether hypothermic CBF/metabolism uncoupling exists on a local cerebral level. This study investigated the effects of hypothermic anesthesia on local cerebral blood flow (LCBF) and local cerebral glucose utilization (LCGU). Methods Thirty-six rats were anesthetized with isoflurane (1 minimum alveolar concentration) and artificially ventilated to maintain normal arterial carbon dioxide partial pressure (pH-stat). Pericranial temperature was maintained as normothermic (37.5 degrees C, n = 12) or was reduced to 35 degrees C (n = 12) or 32 degrees C (n = 12). Pericranial temperature was maintained constant for 60 min until LCBF or LCGU were measured by autoradiography. Twelve conscious rats served as normothermic controls. Results Compared with conscious animals, mean CBF remained unchanged during normothermic anesthesia. Mean CBF significantly increased during mild hypothermia but was unchanged during moderate hypothermia. During normothermic anesthesia, mean CGU was 45% lower than in conscious controls (P < 0.05). No further CGU reduction was found during mild hypothermia, whereas CGU further decreased during moderate hypothermia (48%; P < 0.05). Local analysis showed a linear LCBF/LCGU relationship in conscious (r = 0.94) and anesthetized (r = 0.94) normothermic animals, as well as in both hypothermic groups (35 degrees C: r = 0.92; 32 degrees C: r = 0.95; P < 0.05). The LCBF-to-LCGU ratio increased from 1.4 (conscious controls) to 2.4 (normothermic isoflurane) and 3.6 ml/micromol (mild and moderate hypothermia, P < 0.05). Conclusions Decrease of mean CGU at unchanged or increased mean CBF during hypothermic anesthesia may not indicate uncoupling. Local analysis shows a maintained linear relationship that is reset to a higher CBF/CGU ratio.