Dr Patrick Tsang, Tuen Mun Hospital (A&IC)
Background: Anaemia is common among the critically ill. Previous clinical trials have demonstrated that a restrictive transfusion strategy targeting haemoglobin (Hb) concentration of 7-9g/dl is safe in critically ill patients. Nevertheless, some clinicians advocate a liberal blood transfusion strategy aiming at Hb>10g/dl for those with neurocritical illness such as aneurysmal subarachnoid haemorrhage (aSAH), who may be sensitive to the deleterious effects of anaemia should vasospasm and delayed ischaemic neurological deficit occur. While the optimal transfusion trigger is still debatable, a recent study has suggested an appropriate transfusion threshold in aSAH may be affected by the presence of vasospasm.
Objectives: To assess whether the presence of anaemia or blood transfusion is associated with a worse outcome in patients with aSAH and to assess whether the presence of vasospasm affects the association between anaemia or blood transfusion and outcomes in patients with aSAH.
Design and Setting: Retrospective study in a mixed surgical and medical ICU unit in a tertiary hospital in Hong Kong
Patients and Measurements: Medical records of patients admitted to our unit between January 2005 and January 2009 with the principal diagnosis of aSAH were reviewed. Patients’ demographics, neurological status on admission, initial finding on computer tomography (CT) of brain together with laboratory data on Hb level and the number of packed cell units transfused were documented. Clinical outcomes and complications were also recorded.
Results: 98 patients with median age of 54.5years old were recruited. Anaemia (nadir Hb<10g/dl) was present in 58 patients (59.2%) while 31 patients (31.6%) had received blood transfusion. After adjusting for potential confounders with multiple logistic regression, both anaemia and blood transfusions were independent predictors of i)poor outcomes defined as the presence of in hospital death, vegetative state or severe disability (odd ratio [OR] for anaemia, 2.8; 95% confidence interval [CI] 1.2-6.6; p=0.02; OR for transfusion, 3.4; 95% CI, 1.5-8.5; p=0.01) and ii)combined outcomes of Glasgow Outcome Scale 1-3 and delayed infarctions (OR for anaemia, 2.4; 95% CI 1.0-5.7; p=0.04; OR for transfusion, 4.5; 95% CI, 1.5-13.1; p=0.01). The association between anaemia and poor outcomes was stronger in those without vasospasm while transfusion was a stronger predictor of poor outcomes in those with vasospasm.
Conclusions: Anaemia or transfusion was independently and significantly associated with poor outcomes in patients with aSAH. In the presence of vasospasm, blood transfusion was not associated with improved outcomes. Therefore, the use of vasospasm as a transfusion trigger cannot be justified in this study. Instead, its presence supported a restrictive blood transfusion strategy which may improve cerebral blood flow and cerebral oxygen delivery. Further randomized trials on different transfusion triggers and alternative blood conservation techniques in managing ICU anaemia in neurocritical patients are warranted.