|Home | About | Journals | Submit | Contact Us | Français|
Anemia is common after subarachnoid hemorrhage (SAH) and may exacerbate the reduction in oxygen delivery (DO2) underlying delayed cerebral ischemia (DCI). The association between lower hemoglobin and worse outcome, including more cerebral infarcts, supports a role for red blood cell (RBC) transfusion to correct anemia. However, the cerebral response to transfusion remains uncertain, as higher hemoglobin may increase viscosity and further impair cerebral blood flow (CBF) in the setting of vasospasm.
Eight patients with aneurysmal SAH and hemoglobin < 10 g/dl were studied with 15O-PET before and after transfusion of 1 unit of RBCs. Paired t-tests were used to analyze the change in global and regional CBF, oxygen extraction fraction (OEF), and oxygen metabolism (CMRO2) after transfusion. DO2 was calculated from CBF and arterial oxygen content (CaO2). CBF, CMRO2 and DO2 are reported in ml/100g/min.
Transfusion resulted in a 15% rise in hemoglobin (8.7±0.8 to 10.0±1.0 g/dl) and CaO2 (11.8±1.0 to 13.6±1.1 ml/dL, both p < 0.001). Global CBF remained stable (40.5±8.1 to 41.6±9.9), resulting in an 18% rise in DO2 from 4.8±1.1 to 5.7±1.4 (p = 0.017). This was associated with a fall in OEF from 0.49±0.11 to 0.41±0.11 (p = 0.11) and stable CMRO2. Rise in DO2 was greater (28%) in regions with oligemia (low DO2 and OEF≥0.5) at baseline, but was attenuated (10%) within territories exhibiting angiographic vasospasm, where CBF fell 7%.
Transfusion of RBCs to anemic patients with SAH resulted in a significant rise in cerebral DO2 without lowering global CBF. This was associated with reduced OEF, which may improve tolerance of vulnerable brain regions to further impairments of CBF. Further studies are needed to confirm the benefit of transfusion on DCI and balance this against potential systemic and cerebral risks.
Delayed cerebral ischemia (DCI) is the principal cause of secondary brain injury following aneurysmal subarachnoid hemorrhage (SAH).1 Several factors converge to impair cerebral blood flow (CBF) and oxygen delivery (DO2),2 including vasospasm, hypovolemia, and failure of cerebral autoregulation.3 The cerebral circulation compensates for this reduction in DO2 by increasing oxygen extraction fraction (OEF),4 so that the total amount of oxygen available for cellular metabolism (DO2 × OEF) is maintained, and ischemia, with a fall in cerebral metabolic rate of oxygen (CMRO2), is averted. Tissue in this precarious state of raised OEF, termed oligemia,5 may be threatened by further reductions in DO2 (e.g. worsening vasospasm) beyond the ability of increased extraction to compensate. Beyond this threshold, CMRO2 becomes compromised, neurological deficits may develop, and such ischemic tissue will progress to infarction if DO2 is not promptly restored. Avoiding critical reductions in DO2, therefore, is central to any therapy that aims to minimize morbidity from DCI.
Cerebral DO2 is determined by both CBF and arterial oxygen content (CaO2), with the latter primarily determined by hemoglobin levels. Therefore, anemia may further impair DO2 to brain regions with reduced CBF, and could promote ischemia. Not only is anemia extremely common after SAH,6 but it has been consistently associated with worse outcome, including an increased rate of cerebral infarction.7-9 However, higher hemoglobin may increase blood viscosity and, combined with autoregulatory vasoconstriction in response to increased CaO2, further reduce CBF, countering any benefit on DO2.10,11 The traditional management of DCI has favored a paradigm of hemodilution,12 lowering hemoglobin in an attempt to augment CBF. As the fundamental objective in managing DCI is maximizing DO2 not CBF, hemodilution, by reducing CaO2, may actually be detrimental to this goal.13
This has generated interest in the ability of red blood cell (RBC) transfusion, by increasing hemoglobin and CaO2, to improve DO2 and protect against DCI. However, transfusion of stored RBCs may confer additional risks beyond increased viscosity, including an increased risk of vasospasm in one retrospective study.14 Therefore, the management of anemia after SAH requires a balancing of risks and benefits, with the optimal hemoglobin level and role of transfusion in these patients remaining unclear.
We studied the cerebral vascular and metabolic response to raising hemoglobin with RBC transfusion using positron emission tomography (PET) in a series of anemic patients after SAH who were at risk for DCI. We tested the hypothesis that transfusion would increase cerebral DO2 in order to determine whether the metabolic state of brain regions exhibiting oligemia and supplied by vessels with vasospasm would improve after transfusion.
Patients were eligible for this study if they: 1) had suffered a spontaneous SAH; 2) had a ruptured cerebral aneurysm secured by endovascular or surgical means; 3) had hemoglobin < 10 g/dL; and 4) were at risk for DCI based on ischemic neurologic deficits, angiographic vasospasm, or admission CT grade.15 Exclusion criteria included active congestive heart failure, pregnancy, or inability to obtain matched blood. Informed consent was obtained from patients or their surrogates. The Human Research Protection Office and Radioactive Drug Research Committee of Washington University approved the study protocol.
All patients with SAH were cared for in the Neurology/Neurosurgery Intensive Care Unit (NNICU). Patients received nimodipine and a 3-day course of anticonvulsants. Ruptured aneurysms were treated within 24 hours of admission. Patients were maintained in a euvolemic state by adjustment of intravenous fluids. New or worsening neurological deficits were promptly evaluated, and if no alternative cause was identified, patients underwent cerebral angiography and hemodynamic augmentation with fluids and induced hypertension.16 Many received endovascular interventions for vasospasm including angioplasty and/or intra-arterial vasodilators. Anemia was generally not treated until hemoglobin fell below 7 g/dl, although some patients were transfused if hemoglobin was < 10 g/dl in the presence of angiographic or symptomatic vasospasm.
Data collected on each patient included demographics, medical and social history, and neurologic status at the time of admission.17 Admission CT was rated using the Fisher scale,15 and the amount of intraventricular hemorrhage (IVH) was measured.18 Daily hemoglobin levels and all RBC transfusions were recorded. Cerebral angiograms were reviewed for the presence of arterial vasospasm, graded as mild, moderate, or severe in each vascular territory.
All patients were studied on the Siemens/CTI ECAT EXACT HR+ PET Scanner located in the NNICU.19 Image acquisition was performed as detailed previously to measure CBF, CBV, OEF, and CMRO2.20 After the first series of scans, a single unit of RBCs (volume 350 ml) was transfused over one hour in the PET facility. Once the transfusion was complete, scans were repeated to obtain post-transfusion images. A NNICU physician was present throughout the study and all ongoing therapies for DCI, including fluids and vasopressors, were continued. At the time of each image acquisition, physiologic data were recorded including central venous pressure (CVP) if available. Subsequent management of anemia and vasospasm was at the discretion of the clinical team.
All PET scans for each patient were co-registered and aligned to the initial baseline CBF study using Automated Image Registration software (AIR, Roger Woods, University of California, Los Angeles).21 These images were then co-registered to a reference brain image and resliced so that data could be localized in Talairach atlas space. The patient's CT scan in closest temporal proximity to the PET study was also realigned with the PET images. Using the individual CT images and brain atlas coordinates as guides, an image mask was created that included the brain below the superior sagittal sinus down to the level of the pineal gland. This was used to measure global values for each parameter before and after transfusion.
Spherical regions of 10-mm diameter were placed in 36 predetermined locations covering the major vascular territories bilaterally, as previously outlined.22 CT images were reviewed and regions corresponding to hematoma, infarcted tissue, or ventricular system were excluded. Regional values were then calculated within each of the remaining spheres.
Cerebral DO2, the product of CBF and CaO2, was considered low below 4.5 ml/100g/min (equivalent to CBF of 25 ml/100g/min at low-normal CaO2 of 18 ml/dl). Total oxygen extracted was calculated as the product of CaO2 and OEF (AVDO2). Regions with oligemia were defined by low DO2 associated with OEF ≥ 0.5. These thresholds are conservative estimates guided by data from normal controls,23,24 and previous PET studies of SAH patients.4,25
Global values before and after transfusion were compared using Wilcoxon signed ranks tests while regional values were compared using paired t-tests. Response in regions supplied by vessels with and without moderate to severe angiographic vasospasm was compared using one-way ANOVA. A differential response in regions with oligemia was also similarly examined. We searched for any regions where delivery fell (≥ 10%) after transfusion to evaluate the potential for increased hemoglobin to significantly impair DO2.
Eight patients with SAH were studied an average of 8 days after aneurysm rupture (Table 1). All were at high risk for DCI, based on the presence of angiographic vasospasm, thick cisternal clot and/or intraventricular blood. Three patients were being actively treated for ischemic neurological deficits (decreased or altered mental status in all three, focal deficits in one), all associated with severe bilateral arterial vasospasm.
Four had received a RBC transfusion prior, but none within 24 hours of the PET study. No patients were on sedative medications at the time of study. 15O-labeled oxygen studies in two patients were unusable due to technical problems, limiting OEF and CMRO2 data to only six patients. However, measurements of CBF and DO2 were available for all patients.
RBC transfusion resulted in a 15% rise in both hemoglobin and CaO2 (Table 2). There was also a small rise in MAP and trend to higher CVP after transfusion. No transfusion reactions, changes in temperature, or alterations in neurological status were observed after transfusion. Mean storage duration of transfused RBCs was 26 ± 12 days.
Mean global CBF was unchanged after transfusion (Figure 1), resulting in DO2 rising by almost 20% on average. This increase in DO2 was associated with a fall in OEF and a stable CMRO2. Mean global CBV was unchanged (3.8 ± 0.3 to 3.7 ± 0.3 ml/100g), as was AVDO2 (5.9 ± 1.9 to 5.7 ± 1.7 ml/dl).
Of 288 regions in 8 patients, 25 were excluded. The range of the regional DO2 values for each patient before and after transfusion is plotted in Figure 2. There was a correlation between baseline DO2 and OEF (r = 0.46). The cutoff for the lowest quartile of DO2 was 4.2 ml/100g/min. This group had the highest mean regional OEF at 0.57, compared to 0.39 in those in the highest quartile (p<0.001). 84% of all regions had improved DO2, while only 7% manifested a significant reduction after transfusion. The majority of regions (14 of 18) with such a drop in DO2 were found in a single patient with vasospasm but no oligemia. Mean regional OEF also fell significantly (0.52 ± 0.13 to 0.41 ± 0.13, p < 0.001).
Low delivery (< 4.5 ml/100g/min) at baseline was present in 89 regions (34%). These had higher baseline OEF (0.56 vs. 0.45) and lower CMRO2 (2.0 vs. 2.6, both p < 0.001) compared to other regions. DO2 increased by 25% in these at-risk regions and OEF fell by 0.12 (compared to 0.09 in regions with normal DO2 at baseline, p=0.05). 88% of regions with low DO2 had an improvement after transfusion, with number of regions still having low DO2 decreasing by 50% after transfusion.
There were 56 regions with oligemia (mean baseline DO2 3.7 ± 0.6, OEF 0.62 ± 0.1). Improvement in DO2 was greater compared to non-oligemic regions (28% vs. 15%, p < 0.001) and fall in OEF was larger (-0.14 vs. -0.08, p < 0.001, Figure 3A). Almost all such regions (95%) demonstrated improved DO2 after transfusion, and while CBF was unchanged in non-oligemic regions, it actually rose by 11% in oligemic regions (p<0.001). Only 15 of these 56 regions still met criteria for oligemia and no previously normal regions became oligemic after transfusion. Despite the significant improvement in DO2 to oligemic regions, we did not identify any rise in CMRO2.
Four patients had vasospasm demonstrated on angiography (a median of 1.5 days prior to PET), resulting in 86 regions falling within affected territories. These regions had significantly lower baseline CBF (38.4 ± 11.3 vs. 47.8 ± 11.7, p < 0.001), DO2 (4.7 ± 1.3 vs. 5.5 ± 1.5, p<0.001), and higher OEF (0.51 ± 0.18 vs. 0.47 ± 0.13, p=0.04) than other regions. Low DO2 was present in 47% of regions with vasospasm compared to 28% of those without (p=0.002). Oligemia was present in 28% of affected regions compared to 22% of those without vasospasm (p=0.32).
There was a smaller increase in DO2 after transfusion in these regions (10% vs. 24%, p < 0.001) although reduction in OEF was similar (Figure 3B). This corresponded to a significant (7%) drop in CBF within these regions. Even restricting this analysis to only the patients with vasospasm, affected regions had lower CBF and DO2, and an attenuated response to transfusion compared to those regions located outside affected territories. Only 57% of oligemic regions within territories of vasospasm resolved after transfusion, compared to 85% of similar at-risk regions not affected by angiographic vasospasm (p=0.02).
In this series of anemic patients at risk for cerebral ischemia after SAH, we have shown that RBC transfusion significantly increased cerebral oxygen delivery throughout the brain. This improvement was greatest in oligemic regions at highest risk for ischemia. The number of brain regions with low delivery was reduced by half after transfusion. This novel finding is clinically relevant, as maximizing oxygen delivery is the cornerstone of preventing neurological injury from DCI. Although this goal is typically achieved by maneuvers intended to augment CBF, our results show that raising hemoglobin can provide similar benefits. Not only did transfusion improve CaO2, but this was not at the expense of reduced CBF, as some have feared. This suggests that any increase in viscosity resulting from raising hemoglobin (at least to these levels) does not adversely affect CBF. In addition, it appears that following SAH, cerebral vessels do not vasoconstrict to lower CBF in compensation for higher CaO2.10 An analogous impairment of pressure autoregulatory capacity underlies the ability of hypertensive therapy to augment CBF after SAH in areas of reduced perfusion.
Transfusion did result in a small drop in CBF in territories affected by arterial vasospasm. While net DO2 still improved, the benefit was attenuated. It is possible that the rise in viscosity at higher hemoglobin levels might be more significant within abnormal vessels, where flow is already reduced and compensatory vasodilatation is not possible. Transfusion may not be as effective in reversing hypoperfusion-induced reductions in DO2 related to vasospasm as it is in correcting oligemia related to anemia in regions with preserved vascular tone. This hypothesis is preliminary, based on a few patients, but deserves further study.
The brain compensates for reductions in CBF and DO2 through vasodilatation and by increasing the fraction of oxygen extracted (OEF), as seen in the setting of unilateral carotid occlusion,26 as well as cerebral vasospasm.4 This response maintains oxidative metabolism at levels adequate to prevent ischemia, but can be exhausted if DO2 falls further. Abnormal vessels may not be able to vasodilate to preserve flow in the setting of vasospasm or anemia,22 and so increasing OEF is of central importance to averting cerebral ischemia. We demonstrated that transfusion reversed the elevated OEF seen in these patients. Interventions that reduce tissue OEF should restore critical compensatory reserve and improve tolerance to further ischemic insults. The ability to prevent progression of hemodynamic compromise to frank ischemia is the central goal of managing DCI.
Our findings, coupled with the strong association between anemia and poor outcome in patients with SAH, suggest that RBC transfusion may be a rational approach to minimizing the burden of DCI. However, enthusiasm for the use of RBC transfusion has been tempered by a number of increasingly recognized risks in critically ill patients, including increased rates of nosocomial infections and organ dysfunction.27 Much of this has been attributed to the deleterious effects of storage of RBCs, which leads to morphological and biochemical alterations that may also limit their ability to supply oxygen to tissues. The P50 of transfused blood is reduced, as 2,3-diphosphoglycerate (2,3-DPG) is depleted over time, impairing oxygen offloading.28 Stored RBCs change shape and become less deformable, which may impair their passage through the microcirculation.29
The clinical implications of these changes are uncertain; transfused RBCs are diluted within a much larger pool of endogenous erythrocytes,30 and 2,3-DPG is rapidly regenerated in vivo.31 We did not detect any drop in CBF in our patients after transfusion to corroborate concerns over reduced flow, except in areas of vasospasm. It is also unlikely that the observed reduction in OEF was explained by impaired oxygen offloading (i.e. a left-shift in the oxyhemoglobin dissociation curve), as the net amount of oxygen extracted (AVDO2) remained stable. This issue could be further explored by comparing the efficacy of fresh vs. stored blood on cerebral oxygen transport and metabolism
There is also specific concern about the use of RBC transfusion in SAH, as one retrospective study found an association between post-operative transfusion and a higher rate of angiographic vasospasm.14 While a more recent study could not confirm this finding, it did show that transfusion was more strongly associated with death, disability, and delayed infarction than anemia itself.9 Whether transfusion directly contributes to poor outcomes, or is simply a marker for more severe SAH, remains unclear from these studies but bears further evaluation.
The optimal hemoglobin level for SAH patients at risk for cerebral ischemia remains uncertain. We have shown a beneficial effect on oxygen delivery when transfusing patients with severe anemia (Hb < 10 g/dl). Yet after transfusion all still remained anemic and many had brain regions with persistently low DO2. However, at higher hemoglobin levels increased blood viscosity might impair CBF sufficiently to counter any benefits of higher CaO2 on net DO2. This might be especially the case in patients and regions affected by vasospasm. Further study is needed to ascertain whether such a hemoglobin ceiling exists above which delivery cannot be further optimized.
Our study has a number of limitations. These findings elucidate the acute response to transfusion, while the sustained effects of raising hemoglobin remain unclear. There may have been a small volume-mediated effect of transfusing blood rapidly to these patients; however, it is unlikely that the small increase in MAP (3 mm Hg) and the trend to higher CVP sufficiently accounts for the large improvement in DO2 that was observed after transfusion. All our subjects were female, largely related to the natural history of SAH and their lower baseline hemoglobin placing them at higher risk for anemia; we are planning on studying male SAH patients in future studies at varying hemoglobin levels. We cannot conclude that 10 g/dl is the optimal hemoglobin threshold for transfusion in these patients or whether transfusing to higher levels would further improve DO2. Finally, how this improvement translates into the ability to minimize ischemia, infarction, and neurologic morbidity remains to be determined.
Until further physiologic and clinical studies outline the benefits and exact role of transfusion in these patients, a liberal transfusion strategy for all patients with SAH cannot be recommended based on our findings. However, we have shown that transfusing anemic patients can reverse the impairments in oxygen delivery and compensatory increases in OEF that are seen after SAH, and that this strategy shows promise in minimizing cerebral ischemia in high-risk patients. Large randomized studies are needed to test the benefits of RBC transfusion on clinical outcomes, and determine the balance between the systemic and cerebral risks of transfusion and the benefits we have begun to elucidate here.
We would like to thank Lennis Lich, Angela Shackelford RN, and the cyclotron and ICU staff for their invaluable assistance in performing this research and caring for these complex patients. We would like to thank Dr. William Powers for reviewing the manuscript and providing valuable feedback.
Financial Support: Barnes-Jewish Hospital Foundation 00956-0807-01 and NIH/NINDS 5P50NS035966-10
Conflicts of Interest: The authors have no relationships with companies or entities whose products or services are related to this research or the subject matter.