Anaemia and red cell transfusion in the critically ill

Given time, almost all patients admitted to an intensive care unit (ICU) will become anaemic. There are several reasons for this. It may be a consequence of the reason for admission, such as following major trauma, haematemesis or intra-operative haemorrhage. There may be continuing blood loss from gastro-intestinal erosions or from drains and wounds. This may be exacerbated by abnormal coagulation which may be pathological or pharmacological. Extracorporeal renal support and blood sampling result in a measurable and␣clinically relevant loss [1].␣Poor nutritional intake, substrate deficiencies and depressed bone marrow function decrease production of blood constituents. In critically ill patients, there is inappropriately low erythropoietin (EPO) secretion and function [2]. Renal and hepatic failure will exacerbate this process. In addition, red blood cell survival may be reduced. The limits of acute anaemia have been widely investigated [3-8] and there is considerable clinical evidence from Jehovah's Witness patients [9-12]. One report describes a patient surviving with a haemoglobin (Hb) concentration as low as 1.8 g.dl−1[13]. Clinical experience suggests that a Hb < 5 g.l−1 is associated with a high mortality [14]. Below this value, mortality is 27% for patients less than 50 years old and 53% for those who are older. In young healthy volunteers undergoing isovolaemic haemodilution, Hb concentrations can fall to < 5 g.dl−1 without any obvious side-effects [15, 16]. The risks are greater in those with compromised coronary blood flow [17]; they appear to need a higher Hb value of at least 6–7 g.dl−1. Although even quite profound anaemia may be well tolerated, some groups of patients appear to benefit from higher haemoglobin levels. Patients with chronic renal failure feel and perform better with a Hb level > 10 g.dl−1[18, 19]. Furthermore correction of anaemia has a medium to long-term benefit in patients with severe congestive cardiac failure [20] and in elderly patients following an acute myocardial infarction [21]. Global oxygen delivery depends upon the Hb, oxygen saturation, dissolved oxygen and cardiac output. In some groups of critically ill patients, studies have shown that survival is improved with a raised oxygen delivery [22], and more specifically with a higher haematocrit, blood volume and red cell mass [23]. Red cells are transfused to increase the oxygen content of blood, improve oxygen delivery and thereby prevent tissue hypoxia. Whilst stored blood will increase global oxygen delivery it may be less effective at the microcirculatory level [24]. In the absence of reliable, clinically usable, direct measurements of the adequacy of tissue oxygenation, the haemoglobin concentration is commonly used as a surrogate marker. In the UK, about 50% of ICU patients are transfused using, in total, 6% of the red cells issued to hospitals. There have been several guidelines for the use of blood products, particularly red blood cells. The most recent from the Association of Anaesthetists [25] suggests that a Hb concentration between 8 and 10 g.dl−1 is safe, even for patients with significant cardiorespiratory disease. There are real concerns that limit the use of blood products in the ICU. The commonest cause of mortality attributed to homologueous blood transfusion is mismatched blood [26]. Transfusion-related acute lung injury (TRALI) is the second most common cause of transfusion-related death. In addition to these severe hazards, there are more subtle detrimental effects associated with a blood transfusion. Stored blood contains inflammatory mediators, largely produced by white cells. These cytokines may affect immune function [27, 28]. Transfusion-related immunomodulation can worsen the altered immune status which accompanies critical illness. This may explain the increased predisposition to postoperative infection [29, 30], increased mortality [31, 32] and recurrence of cancer [33] observed in some groups of patients receiving peri-operative blood transfusions. Leucodepletion may reduce the incidence of some of these adverse effects [32, 34]. Blood products such as albumin and clotting factor concentrates are manufactured by pooling plasma from 20 000 or more blood donations. Because of the theoretical risk of prion transmission, from December 1998, plasma for the manufacture of fractionated products in England has been sourced from the USA. For several years, all blood donations have been tested for Hepatitis C antibodies. From July 1999, Hepatitis C virus nucleic acid testing (NAT) has been used as a release criterion for frozen components such as FFP and cryoprecipitate. From November 1999, blood has been leucodepleted. All these measures to deliver a ‘safe’ blood supply have considerably increased the cost of a unit of blood [35]. Future likely developments, for example screening donors for new variant CJD, have the potential to increase costs further and markedly decrease the number of people willing to donate blood. Therefore strategies to ensure that blood components are used appropriately are of the utmost importance. In situations associated with massive blood loss, transfusion of blood is life saving. However, in more stable groups of patients, when blood may be used to prevent the adverse effects of anaemia, the advantages of achieving a higher Hb concentration have to be balanced against the risk inherent in stored allogenic blood [36]. Hebert conducted a survey into Canadian intensive care transfusion attitudes [37] and found there was considerable variation in transfusion practice [38]. When British intensivists were given ICU patient scenarios and asked to indicate the Hb thresholds they would use for transfusion, few would let the Hb fall below 8 g.dl−1 before transfusing or give blood for a Hb above 10 g.dl−1[39]. When actual practice was examined, a similar pattern was recorded with blood rarely given for Hb concentrations more than 10 g.dl−1[40]. Interestingly, 72% of␣transfusions were prescribed merely because the patient had a low Hb. The appropriate Hb value in critically ill patients has been the subject of a prospective study. Patients were randomly allocated to be transfused at a restrictive threshold (7 g.dl−1) or a liberal threshold (10 g.dl−1) [41, 42]. Although, as with any study of this size and complexity, criticisms can be made of the protocol and results, there was no evidence of better outcome in patients who had the more liberal transfusion policy. Indeed, some subgroups assigned to the restrictive policy had a statistically significant improved outcome. In a further analysis of patients with cardiovascular disease, there was no obvious or statistically significant difference in survival between the groups [43]. One group singled out for benefit from a higher Hb is patients with chronic obstructive pulmonary disease (COPD) who are difficult to wean from a ventilator [44]. A series of case reports suggested that increasing the Hb concentration improved the likelihood that the patient could be weaned [45]. This has not been confirmed in a sub group analysis of the critical care patients randomised to a restrictive or liberal transfusion policy [46]. Blood is a scarce commodity, expensive and potentially dangerous. Blood loss should therefore be minimised. One important cause of anaemia in critically ill patients is the amount of blood taken for investigations [1, 47, 48]. On average, each patient is bled some 50–60 ml.day−1. Typically, this results in an average total ICU blood loss of 500–700 ml, although individual patients can lose much more than this. Using blood conservation devices to sample blood from arterial lines will approximately halve the volume of blood taken [49]. A similar effect will be achieved if paediatric bottles are used for the tests [50, 51]. Pharmacological agents to decrease haemorrhage may be useful in some patients. These drugs include tranexamic acid, aprotonin, and fibrinogen glue. Surgical and anaesthetic techniques can also affect the likelihood and amount of haemorrhage. Autologous blood may be safer and requires predonation or normovolaemic haemodilution. Shed blood can be captured and re-transfused with cell savers. There appears to be a blunted erythropoietic response to anaemia in critically ill patients [52-54], and administering additional EPO decreases the requirement for blood transfusion [55]. Whilst liberal transfusion of allogenic red cells is associated with a worse outcome in some critically ill patients [41], higher haemoglobin levels, when achieved with EPO, improve quality of life in patients with chronic renal failure [18, 19] and are associated with a better outcome in patients with serious congestive cardiac failure [20]. Thus, at least in stable groups of patients, EPO should be considered as an alternative to red cells. The observation by Rao et al. [40], that the indication for transfusion of red cells was a low haemoglobin level in the majority of patients, supports this view. Although EPO is expensive, the increase in the cost of blood will make it more cost effective and therefore worthy of further investigation. If, despite everything, a patient becomes anaemic, the real and potential harm of a blood transfusion should be balanced against the evidence indicating that a Hb concentration down to 8␣g.dl−1, or even lower, is safe in the critically ill.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.