Introduction

Aneurysm rupture causing subarachnoid hemorrhage (SAH) occurs in about 10/100,000 people each year [1, 2]. Nearly half of these individuals are dead within 30 days [1–5]. Among survivors, only one-third make a full recovery and approximately half who appear to experience a favorable outcome have neuropsychological and cognitive deficits and difficulties in their daily activities [6–10]. Poor outcome after SAH may be associated with two preventable factors delayed cerebral ischemia (DCI) and extracerebral organ dysfunction (e.g., medical complications and infection) [11–19]. The mortality rate from extracerebral organ dysfunction is 20–40% [17, 18]. DCI occurs in about 30% of patients, is often associated with arterial vasospasm that begins about 3 days after SAH, is maximal days 7–8, resolves after 14 days, and is identified radiographically in about 70% of patients [20–24]. Clinical trials to prevent vasospasm seldom have improved clinical outcome, despite reduced vessel narrowing [21, 25, 26]. This dissociation between clinical outcome and vasospasm has refocused efforts to limit brain injury rather than vessel narrowing and has renewed interest in intensive care strategies to prevent DCI and medical complications.

DCI is caused by impaired cerebral blood flow (CBF) and O2 delivery (DO2). Cerebral circulation compensates for reduced CBF by increasing oxygen extraction fraction (OEF) to maintain the amount of oxygen available for metabolism and to prevent ischemia. When OEF is increased (oligemia), tissues may not compensate for further DO2 reductions and, if not corrected, infarction may occur. Therefore, avoiding critical DO2 reductions is central to SAH care. CBF and CaO2 determine cerebral DO2, and since hemoglobin (Hb) levels primarily determine CaO2, anemia may impair cerebral DO2. After SAH, more than half the patients develop anemia [27, 28] that maybe associated with worse outcome. Clinical studies also suggest that Hb > 11 g/dL may be associated with improved SAH outcome [29–31]. However, higher Hb may increase blood viscosity and, with autoregulatory vasoconstriction in response to increased CaO2, further reduce CBF, countering any DO2 benefit. Furthermore, RBCT (red blood cell transfusion) has been associated with organ dysfunction and mortality [32–35]. This effect may be mediated by inflammatory mediators or altered nitric oxide (NO) metabolism among other factors in transfused cells [36–41]. Both inflammation and NO influence vasospasm [21, 42]. Limited data are available to guide anemia management and RBCT after SAH.

This literature review was designed to present available published evidence on the occurrence and outcome of anemia in SAH patients. The role of RBCT and erythropoietin on brain physiology and clinical outcome also was explored.

Methods

A literature search was made to identify clinical or experimental studies published between 1980 and August 2010 in the English literature that described and compared RBCT strategies and Hb levels after rupture of a cerebral aneurysm. Candidate articles were identified from electronic databases, including Medline and EMBASE, Index Medicus, bibliographies of pertinent articles, and expert consultation. Additional articles were identified through review of textbooks, bibliographies from retrieved articles, and the “Related Articles” feature of PubMed. For electronic searches, the following key words were used: “subarachnoid hemorrhage,” “subarachnoid hemorrhage outcome,” “anemia,” “hemoglobin,” “transfusion,” “packed red blood cells,” “vasospasm,” “delayed cerebral ischemia,” “blood products,” and “erythropoietin.” This search was supplemented by also identifying randomized trials that have compared transfusion strategies in general critical care, recent review articles on transfusion in neurocritical care, transfusion guidelines, and studies that have evaluated transfusion and hemoglobin in traumatic brain injury (TBI).

Original research studies were selected for detailed review if they addressed incidence and/or outcome of anemia and treatment with RBCT or erythropoietin after SAH. Selected studies were evaluated for quality of evidence using the GRADE system [43].

Summary of the Literature

Five hundred and twelve manuscripts were identified. There is no high-quality evidence to support a particular transfusion strategy or Hb level in patients with SAH. Twenty-seven articles were selected for detailed review (Tables 1, 2, 3, 4) [27, 29–31, 44–66]. These studies addressed brain physiology related to transfusion or Hb (11 articles—5 in traumatic brain injury and 6 in SAH; Table 1), anemia in SAH (5 articles; Table 2), RBCT management in SAH (8 articles; Table 3), and erythropoietin after aneurysm rupture (4 articles; Table 4). Most studies describe retrospective analyses of clinical series. Overall, the quality of evidence is very low according to the GRADE criteria. One small, randomized pilot study evaluated two transfusion strategies in SAH [58]; however, this study was underpowered, and no conclusions about a particular strategy could be made. The study, however, suggests that a randomized trial is safe and feasible.

Table 1 Summary of published literature that examines the effect of RBCT on brain oxygen and metabolism
Full size table
Table 2 Summary of published literature that examines anemia and SAH
Full size table
Table 3 Summary of published literature that examines the clinical effects of RBCT in SAH
Full size table
Table 4 Summary of published literature that examines EPO and SAH
Full size table

The data summarized in the tables support that anemia affects about half of patients with SAH and is linked with worsened outcome. While RBCT has been shown to have beneficial effects on brain physiology, RBCT is associated with medical complications, infection, vasospasm, and poor outcome after SAH. It remains unclear whether RBCTs are simply a marker of disease severity or an independent cause of worse outcome. Erythropoietin data are very limited.

In addition to those articles meeting criteria for detailed review, additional publications provide clinically useful information on the potential role of RBCT in SAH. These studies are summarized below.

Anemia After SAH

Anemia is common after SAH. Depending on the definition applied, anemia has been identified in 40–50% of SAH patients and only 16% maintain Hb > 11 g/dL [27, 54, 55, 67]. The mean drop in Hb after SAH is 3 g/dL, and anemia develops after a mean of 3.5 days [27]. Anemia may exacerbate the reduction in oxygen delivery that underlies DCI. Observational studies have linked anemia or a larger Hb reduction with infarction, dependency, and death after SAH [19, 29, 55, 68]. In addition, patients with an unfavorable outcome consistently have lower Hb levels, especially between days 6 and 11, following SAH (i.e., during the greatest risk period for DCI) [55].

Experimental evidence links anemia with reduced PbtO2 and increased neuron injury after acute brain injury [69–71]. In the normal brain, compensatory vasodilation occurs with Hb < 10 g/dL [72], so brain hypoxia usually is manifest only at lower Hb levels (e.g., <6 g/dL) [73]. When cerebrovascular reserve is impaired, e.g., in patients with SAH, tissue hypoxia and cell injury may develop at a higher Hb. For example, using cerebral microdialysis in 20 poor-grade SAH patients, Hb ≤ 9 g/dL was identified as an independent factor associated with cerebral tissue injury [53]. In a similar study in which patients requiring FiO2 > 60% were excluded, Kurtz et al. linked Hb < 10 g/dL with cell energy dysfunction [54]. Consistent with this, mathematical modeling based on animal experiments of brain ischemia suggests that Hb < 10 g/dL is associated with brain hypoxia [71].

Correction of anemia with RBCT may, therefore, improve PbtO2 and attenuate cell damage. In SAH patients, anemia can be associated with poor outcome, and avoidance of low Hb may, therefore, be warranted [56]. The optimal Hb threshold for RBCT in SAH patients remains unclear although a recent clinical study suggests that Hb > 11 g/dL is associated with less cerebral infarction and improved outcome after SAH [30].

RBCT in General Medical and Surgical Critical Care

In most cases, RBCT is used in critical care for the treatment of anemia [74, 75], with a commonly used Hb cutoff of 10 g/dL to augment oxygen delivery and avoid oxygen debt [76]. This practice is now challenged by evidence that suggests RBCT may exacerbate outcome and increase medical complications in general critical care [32–34]. Consequently, a restrictive RBCT policy (Hb ~ 7 g/dl) may be preferred.

In general critical care patients, RBCT is associated with complications such as immunosuppression, transmission of infectious agents, postoperative infections, and pneumonia [77–83]. RBCT also is an independent risk factor for impaired pulmonary function and prolonged ventilator support, acute lung injury, acute respiratory distress syndrome (ARDS) [84, 85], systemic inflammatory response syndrome [83, 86], renal dysfunction [87], multiple organ failure or dysfunction [34, 88, 89], transfusion reactions [39], and increased length of stay [33]. In SAH patients, RBCT has also been associated with medical complications and infection [29, 57].

Recent observational data suggest that many intensive care patients can tolerate Hb of 7 g/dL, “restrictive” RBCT is safe, or that RBCT may exacerbate outcome or increase complications [35, 74, 80, 89]. These studies, however, included few if any patients with neurological disorders or SAH. There is a dose effect, but as little as one unit of blood may be deleterious [90]. Two randomized trials, one in adults [32] and one in children [91], have addressed RBCT in critical care. The Transfusion Requirements in Critical Care Trial (TRICC) compared a “liberal (10 g/dL)” and “restricted (7 g/dL)” RBCT trigger in 838 ICU patients [32]. Overall 30-day mortality was similar, with lower mortality in the restrictive RBCT group among younger (<55 years) and less ill (APACHE II < 20) patients. Concern remains, however, that restrictive RBCT may not be tolerated in patients with some conditions, e.g., reduced cerebrovascular or cardiac reserve. Consequently, patients with ARDS, sepsis, myocardial ischemia, or traumatic brain injury may require higher Hb levels [92–94]. For example, there may be a benefit to liberal RBCT in elderly patients, with acute coronary syndrome and admission Hb < 8 g/dL [93, 94]. The CRUSADE initiative data from 44,242 patients with non-ST segment elevation acute coronary syndrome suggest that the association between RBCT and outcomes was a function of nadir hematocrit [95]. RBCT tended to have a beneficial impact with hematocrit <25%, with increased mortality when nadir hematocrit was >27%. Conversely, liberal RBCT does not appear to benefit patients who require prolonged mechanical ventilation, where theoretically the oxygen carrying benefit of RBCT might hasten recovery [96, 97]. Together, these various data suggest the effect of RBCT on outcome may depend on the need for oxygen delivery and on the individual patient and his or her pathology.

RBCT in SAH

Retrospective studies report that about one-quarter of patients with SAH receive RBCT during surgery and up to two-thirds during their intensive care stay [31, 98–100]. The first RBCT during intensive care is generally administered a mean of 4.6 days after SAH [31], i.e., just as vasospasm becomes maximal.

There are good theoretical reasons to maintain a higher Hb after brain injury, since the brain has stringent O2 requirements. Most neurosurgeons prefer Hb > 10 g/dL for patients with acute brain injury to maintain optimal oxygen carrying capacity; however, there is variability in the response of brain tissue O2 (PbtO2) to RBCT [44]. Even when RBCT improves PbtO2, it is unclear whether this improves brain metabolism [47]. Few studies have investigated RBCT effects on outcome after brain injury. Subset analysis of 67 traumatic brain injury patients enrolled in the TRICC trial [101] suggests TBI patients can have a similar RBCT threshold to other intensive care patients, while observational data suggest RBCT is associated with worse outcome in this population [102–104]. Many SAH patients, unlike traumatic injury patients, have associated cardiac dysfunction [105], a relative contraindication to restrictive RBCT suggesting that RBCT in SAH needs specific study. PET studies, for example, show that RBCT can improve CaO2 and DO2 without a detrimental effect on CBF in SAH patients with anemia [51].

Data evaluating an effect of RBCT in patients with SAH are limited, despite fluid status and oxygenation being critical to patient care. Not all studies demonstrate an association between RBCT and poor outcome after SAH [59, 61]. However, recent observational studies and post hoc analysis of other trials link liberal use of RBCT with medical complications, infection, vasospasm, poor cognitive performance, and poor outcome [29, 44, 55–57, 60, 68]. It is conceivable that avoiding hypoxia, rather than anemia alone, may prevent neuron damage [106]. RBCT, however, does not always increase PbtO2, and in 20–25% of patients, PbtO2 may decrease [44, 46]. A plausible biologic explanation for the deleterious effects of RBCT is that stored red blood cells have been associated with immunomodulation, impaired vasoregulation, hypercoagulation, altered nitric oxide metabolism, reduced red blood cell deformability, altered red blood cell adhesiveness and aggregability, reduced 2,3-diphosphoglycerate, and impaired microvascular perfusion [36–39]. Some of these factors appear integral to vasospasm [21, 107, 108].

Clinical Management Strategies for Anemia in SAH

A recent international survey queried intensive care physicians about SAH care [109]. Among the 626 respondents, recommendations for optimal Hb ranged from ~8 g/dL (25%) to ~12–13 g/dL (40%). Two-thirds advocated a target Hb > 10 g/dL. There is also widespread variation in the use of RBCT in treating SAH patients, although practices differ from those used in general surgical and medical conditions [110]. Discrepancies identified in clinical practice highlight the need for more research to specifically identify the role of anemia and anemia management in patients with SAH. In addition, there remain many unanswered questions about transfusion after SAH including the role of the following: (1) plasma and platelet component therapies, (2) leukocyte reduction, (3) age of transfused cells, (4) blood product substitutes, and (5) the clinical or biological end point for RBCT. There has been limited study of erythropoietin use, and no firm recommendations about its use can be made.

Conclusions

Anemia develops in about 50% of SAH patients and often within 3 days of aneurysm rupture. Risk factors for anemia after SAH include female sex, advanced age, worse clinical grade, lower admission Hb, and surgery. Anemia has also been identified as a risk factor for poor outcome after SAH. It is not clear whether anemia is an independent factor associated with outcome or a marker of disease severity. However, patients in worse clinical grade or those who develop vasospasm are more likely to have a worse outcome if they develop anemia.

There is limited information about how often SAH patients require RBCT, but recent retrospective studies demonstrate that about one-quarter receive RBCT during surgery and up to two-thirds during their intensive care stay. Physiological studies show increases in brain oxygen in 75% of transfusions and increases of brain DO2. RBCT, however, have been associated with vasospasm, medical complications, infections, worse outcome, and cognitive impairment. When both anemia and RBCT are entered into outcome models, transfusion has a greater effect. Again, it is not clear whether RBCT is a marker of disease severity or an independent risk factor for worse outcome. While the overall quality of literature that examines transfusion in SAH is low, it is clear that the results of the TRICC trial and subsequent observational studies of transfusion in general critical care do not and should not apply to SAH patients. For now, clinicians will need to base transfusion decisions for SAH patients in the context of conflicting information and so should focus on an individualized assessment of anemia tolerance, consider blood conservation strategies, and understand the potential risks and benefits of blood transfusion. Further prospective investigations to address the role of anemia, the optimal Hb threshold, and the use of RBCT in SAH are desperately needed.