(1)
Department of Intensive Care, Guy’s and St. Thomas’ National Health Service Foundation Trust, London, UK
Abstract
Anaemia is extremely common in the critically ill and up to 50% of all patients admitted to critical care will receive a blood transfusion. No blood transfusion can be completely safe and there have been persistent concerns raised about possible increased mortality and morbidity associated with blood transfusion. These concerns prompted a number of studies examining restrictive transfusion practice. More recently, new devices have been developed to minimise blood loss associated with phlebotomy. This chapter will review the literature as regards specifically restrictive transfusion strategies. The later part will focus on therapeutic interventions, which can reduce the amount of blood lost to routine sampling. When a restrictive transfusion strategy is combined with the use of blood conservation devices, there is a trend towards a reduction in transfusion requirements. There is the potential to significantly reduce iatrogenic anaemia and consequently reduce the potential risk of any adverse consequences of transfusion.
10.1 Introduction
According to the SOAP study, 80 % of patients have a haemoglobin concentration of less than 90 g/L 7 days after admission to an intensive care unit (ICU) [1]. This study also examined both transfusion and phlebotomy practices, demonstrating a mean blood loss of 41 mL/day/patient due to phlebotomy. Equating to over 280 mL each week per patient, this volume is almost identical to that of a unit of red blood cells (RBCs). In addition to phlebotomy, haemodilution and haemorrhage are the most significant contributors to the development of anaemia in the critically ill. The high rate of blood loss is compounded by impaired erythropoiesis secondary to inflammation, and high phlebotomy requirements impact increasingly with prolonged critical illness [2]. This extremely high prevalence of anaemia places a significant burden on the national blood services throughout the world with approximately 10 % of all RBCs transfused nationally given in general ICUs [3]. Of note, only about 20 % of these transfused RBCs are to patients who have active bleeding.
A number of publications have linked the transfusion of RBCs to potential harm in the critically ill. Although no definitive pathophysiological process has been proven, it has been demonstrated that blood transfusions are associated with an increased risk of infection and possibly mortality [4]. Despite the complex link between inflammation and the development of anaemia, it is rational to assume that if the volume of blood taken from a patient in routine testing can be reduced, then the development of anaemia may be slowed. This chapter reviews the more recent literature favouring restrictive transfusion policies in the ICU and discusses approaches to blood conservation.
10.2 Incidence of Anaemia and Frequency of Transfusion in Critical Care
Depending on case mix, between 30 and 50 % of patients receive an RBC transfusion during their ICU admission [2, 3, 5]. The majority of these transfusions are given to treat anaemia, with the mean blood volume administered ranging from 2 to 4 units per patient [1]. Despite substantial efforts to make blood safe, no transfusion is risk-free. In addition to the risks of transfusion-transmitted infection, ABO mismatch, volume overload and transfusion-related acute lung injury, the use of blood product support has been consistently associated with poorer patient outcomes.
In the European SOAP study, transfused patients had a longer ICU stay and higher associated ICU mortality [1]. This finding was corroborated by the North American CRIT study, which demonstrated that more transfused RBC units were independently associated with worse clinical outcomes and an admission haemoglobin concentration less than 90 g/L was also correlated with adverse outcomes [6]. These findings are supported by a recent systematic review of 45 studies focusing on patient outcome in relation to transfusion, where RBC use was found to be an independent risk factor for infection and the development of multi-organ failure [7].
10.3 Studies That Have Investigated Restrictive Transfusion Practice
The strongest evidence guiding transfusion policy in adult critically ill patients comes from the transfusion requirements in critical care (TRICC) study [8]. Patients with a haemoglobin concentration of less than 9 g/dL were assigned to one of two arms: either a ‘high’ haemoglobin transfusion trigger of less than 10 g/dL with a target of 10–12 g/dL or a ‘restrictive’ target with a lower transfusion trigger with a haemoglobin concentration less than 7 g/dL. Mortality was compared at 30 and 60 days. The restrictive group received 54 % fewer units of blood with one third receiving no transfusion at all, whereas all the patients in the liberal group were transfused. The 30-day mortality in the liberal group was typical of general ICU populations, but there was a nonsignificant trend towards lower mortality for the restrictive group.
In two predefined subgroups, patients younger than 55 years and patients with an APACHE II score less than 20, the risk of death during the 30-day follow-up was significantly lower with the restrictive strategy. Patients aged less than 55 years who were enrolled in the restrictive arm had a 5.7 % mortality compared to 13.0 % amongst those in the liberal group. Similarly, patients with an APACHE II score less than 20 had an 8.7 % mortality versus 16.1 % when the restrictive and liberal arms were compared. These differences represented a number needed to treat to benefit from restrictive over liberal transfusion of about 13 for both these subgroups. Overall, there were also lower rates of new organ failures in the restrictive group and a trend towards higher rates of acute respiratory distress syndrome in the liberal group (7.7 % versus 11.4 %).
Despite the impressive results demonstrated in the TRICC study, persistent concerns have been raised about the applicability of the results to current practice. The possibility of selection bias has been raised as few patients with cardiac disease were enrolled and there was a high clinician refusal rate to allow randomisation. The study was also performed prior to the mandatory introduction of leucodepletion in many countries. Leucodepletion was introduced in an attempt to mitigate the risk of variant Creutzfeldt-Jakob disease infecting patients through the blood supply chain. Following this, there has been considerable debate as to the consequences of leucodepletion on the safety of blood. There has been a reduction in febrile nonhaemolytic transfusion reactions of around 60 %. CMV transmission has also been significantly reduced and there has been a decrease in the rate of HLA allo-immunisation [9]. The UK haemovigilance system has consistently shown a reduction in the incidence of transfusion-associated lung injury [10].
Despite the concerns of the TRICC study, the results have been supported by a number of other studies. The transfusion requirements after cardiac surgery (TRACS) study found no difference in a composite end point of 30-day mortality and severe comorbidity in cardiac patients prospectively randomised to a liberal or restrictive transfusion strategy [11].
The ‘FOCUS’ study, although not in critically ill patients, compared liberal with restrictive transfusion in high-risk patients undergoing hip surgery. There was no difference in mortality in the group assigned to the restrictive transfusion strategy [12]. It is important to emphasise that although patients in the FOCUS trial were not critically ill, they were elderly and had a high prevalence of cardiovascular disease.
To further study the use of ‘restrictive’ transfusion strategies in critically ill patients over the age of 55, Walsh and colleagues examined the outcome of critically ill patients ventilated for more than 4 days [13]. Patients were randomised to similar ‘liberal’ and ‘restrictive’ groups as previously defined. Baseline comorbidities and illness severity were high, and 32 % of patients had documented ischaemic heart disease. Mortality trended towards a higher rate in the liberal group, and no significant differences were observed in organ dysfunction or duration of ventilation.
A more recent study examined transfusion thresholds in patients with acute gastrointestinal bleeding. Villanueva and colleagues compared restrictive and liberal transfusion strategies in 921 patients [14]. Patients were randomly assigned, with 461 in the restrictive arm compared to 460 treated with a liberal transfusion approach. A total of 225 patients assigned to the restrictive approach as compared to 61 patients following the liberal pathway did not receive transfusion. There was also a reduced incidence of rebleeding, 10 % versus 16 % in the ‘restrictive’ arm. The probability of survival was greater at 6 weeks and the patients suffered fewer adverse events in the restrictive arm. However, when stratified into Child-Pugh score cirrhosis subgroups, there was no significant difference in patients with class C disease, probably reflecting the poor prognosis associated with severe advanced liver disease.
In summary, there is now an increasing body of literature that has studied restrictive transfusion practice with a consistent signal showing that ‘restrictive’ strategies reduce the amount of blood given to patients whilst revealing no clear detrimental effect.
10.4 Blood Sampling in the Critically Ill
Daily blood testing and arterial blood gas draws are the most common investigations performed in the critically ill. It is no surprise that the sickest patients require a higher frequency of blood sampling [15, 16]. Patients with arterial catheters have almost double the frequency of blood tests and a threefold increase in their volume of blood loss, and the total volume of diagnostic blood taken is an independent predictor of transfusion [17]. Arterial blood gases account for around 40 % of blood taken from ventilated patients [18].
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
