Practical Transfusion Medicine 4th Ed.

35. Autologous transfusion

Dafydd Thomas1, Biddy Ridler2 & John Thompson2

1Morriston Hospital, Swansea, Wales, UK

2Peninsula College of Medicine and Dentistry, Royal Devon and Exeter Hospitals, Exeter, UK

Autologous (patient's own) blood transfusion has come full circle. The salvage and reinfusion of blood lost during an amputation was first reported by Dr John Duncan in 1885. The development of blood storage, banking and understanding of serology then led to safe and effective transfusion support for medicine, surgery and obstetrics using donor blood. However, the increased cost of blood components, decreasing numbers of blood donors, increased demand, inappropriate use and successive threats of bacterial, viral and prion infection changed our emphasis at the end of the last century [1–3]. Perhaps the major driver was a medicolegal one, with a new benchmark, namely ‘The public is entitled to expect that the blood they receive will be 100% safe [4]. The knowledge of the medical profession is not relevant in determining the legitimate expectation of the public and nor is the fact that the defect could not have been avoided in relevant circumstance. Once the risk is known about the product is defective even if the risk could not be identified in the particular product’ (Lord Justice Burton, 2001).

As a result of haemovigilance several successful measures were introduced to improve blood safety and the various techniques for the provision of autologous blood have been refined, studied scientifically and audited.

The ‘Appropriate Use of Blood’ subcommittee was established by NHS Blood and Transplant in England to discuss and investigate strategies for blood conservation. It came to the conclusion that an integrated programme would be more successful than piecemeal implementation and that autologous blood transfusion should be part of a total quality management approach based on four strands (Table 35.1):

·        preoperative identification of patients at increased risk of bleeding and optimization before surgery (e.g. correction of anaemia);

·        perioperative blood salvage (collection of blood that would otherwise be lost in the surgical field or in postoperative drains);

·        blood sparing methods such as drugs and surgical technique; and

·        a strict postoperative transfusion protocol.

Table 35.1 An approach to blood conservation and the reduction of risk associated with blood transfusion in patients having elective surgery.

·        Check the blood count well in advance of surgery and correct any treatable anaemia

·        Ask about antiplatelet or anticoagulant drugs the patient is taking and consider if any should be stopped

·        Abide by the maximum/blood ordering schedule that should be available in your organization

·        Check antibody status and blood group so group-specific blood can be used in an emergency

·        Consider whether the patient has a hereditary or acquired bleeding tendency and investigate/treat as appropriate

·        During surgery consider technical methods to reduce bleeding

·        Postoperatively, consider whether blood transfusion is clinically indicated (transfusion trigger) and, if it is, consider how many units are required to achieve the desired Hb (transfusion target)

·        If operation would normally require blood transfusion, consider the option of autologous blood transfusion

 

Such as:

 

Technique

Situations in which it might be considered

Intraoperative cell salvage

Any patient with estimated blood loss >0.5 L; especially suitable for massive blood loss

Postoperative cell salvage

Patients with postoperative drain loss from a clean site

Reasons to consider autologous transfusion

Clinical transfusion practice should reduce the risks involved in blood transfusion. Strict blood donor exclusion criteria together with extensive testing of blood to decrease the risk of transfusion-transmitted infection has reassured clinicians and maintained demand for donor blood. Blood conservation strategies should minimize the use of donor blood by withholding transfusion until strictly clinically necessary and employing techniques such as autologous transfusion. In some situations, autologous transfusion is definitely indicated, such as in patients with rare blood groups or complex red cell antibodies for whom it is difficult to source compatible blood. Autologous transfusion should also be used instead of, or to supplement the use of, donor blood, in situations where it has been shown to be effective and safe. It has been suggested that more than 20% of surgical demand can be met by autologous transfusion [5] and other blood conservation methods such as preoperative anaemia optimization [6]. Certain procedures can be undertaken with virtually no donor blood support (see Chapter 34), conserving supplies for areas of medicine where there are few alternatives, such as haematological oncology and the increased use in upper gastrointestinal haemorrhage.

Blood conservation strategies

Since the third edition of Practical Transfusion Medicine, there has been considerable research, audit and re-evaluation of blood sparing strategies, with the result that several are no longer used in clinical practice.

Pre-deposit autologous donation

This technique was much vaunted as a means of donating and reserving one's own blood prior to elective surgery [7]. Several observational (but very few randomized) studies were published. The problems were as follows:

·        Patients were chronically anaemic at the time of surgery, with little reserve for bleeding and so they were subsequently transfused more frequently.

·        Anaemia increases the bleeding time and therefore surgical blood loss.

·        The practical difficulties were significant both for the patient and the participating hospital.

·        Cancelled operations meant that blood could go out of date.

·        Prospective randomized trials showed comparatively modest savings in donor blood transfusion.

·        Careful blood transfusion protocols in the control groups seriously reduced overall transfusion and surgical teams began to introduce other more effective techniques to reduce blood loss, because they knew they were being observed.

·        There was a lower threshold for transfusing the pre-deposited blood (because it was there), but there was still the potential for clerical or other errors.

Pre-deposit may be useful in certain special situations, such as in paediatric surgery, where there is a great incentive to avoid transfusion-associated infection. In such cases, it can be assisted by stimulating the patient's bone marrow with recombinant erythropoietin and iron (usually parenterally) to increase haemoglobin (Hb) concentration. Directed pre-deposit from relatives was abandoned for ethical reasons as it would place relatives under pressure to reveal lifestyle choices that they might want to keep private.

Acute normovolaemic haemodilution

Acute normovolaemic haemodilution (ANH) seemed to have great promise but is seldom practised outside specialist surgical areas. Blood is withdrawn at the beginning of surgery and replaced with a balance of colloid (usually complexed starch) and clear fluid to maintain normovolaemia. The patient consequently bleeds dilute blood during the procedure, so decreasing red cell loss. After surgical blood loss ceases, the fresh autologous blood is returned. The problems are as follows:

·        There is no level 1 scientific evidence (from a properly powered randomized controlled trial) to show that ANH actually reduces donor blood exposure.

·        ANH takes on average 20 minutes to perform and theatre time is precious.

·        Haemodilution may increase bleeding by reducing the haematocrit and diluting clotting factors.

·        Haemodilution may precipitate cardiac ischaemia and even myocardial infarction.

Most studies of ANH were not randomized or used ANH in conjunction with other methods, so it was difficult to ascribe benefit to one or the other. In addition, although theoretical formulas could be used to determine exactly how much blood could be withdrawn, in clinical practice most authors reported comparatively modest volumes of ANH. There may be circumstances where high volume ANH could be employed in fit patients at low risk of myocardial ischaemia, particularly paediatric spinal surgery. Pilot studies have been successful and randomized trials in niche areas are awaited.

Other techniques that have been abandoned

·        Routine preoperative coagulation tests (unless there is a positive personal or family history of bleeding).

·        Transfusion if Hb >10 g/dL (accepted as unnecessary).

·        Aprotinin (withdrawn by the manufacturer because of renal failure and decreased survival).

·        Ultrafiltration to remove excess water after cardiac bypass.

·        Leucocyte filters in bypass circuits (may activate white cells).

·        Platelet-rich plasmapheresis (no benefit).

·        Unwashed mediastinal blood (may lead to coagulopathy).

·        DDAVP (unless there is an established platelet defect such as uraemia or von Willebrand's disease).

·        Bovine thrombin-derived haemostatic sealants in cardiac surgery (can provoke antibody response and allergy), although the new generation selected male donor products are less likely to cause these problems.

Effective methods for blood conservation

The following techniques are accepted and should be part of a hospital's blood conservation strategy:

·        Total quality management – continuous audit and improvement of the process of blood transfusion in the perioperative setting (see Chapters 27 and 34).

·        Intraoperative cell salvage (ICS).

·        Postoperative cell salvage (PCS) from wound drains.

·        Appropriate transfusion and acceptance of the very low risk of viral and other risks.

·        Transfusion if Hb < 7 g/dL postoperatively.

·        Stopping drugs associated with increased bleeding such as aspirin (unless a high cardiac risk) or clopidogrel (unless a drug eluting coronary stent was inserted in the last 6 months).

·        Blood component therapy if active oozing and supported by abnormal tests and thromboelastography.

·        Near patient Hb testing.

·        Limited sampling in intensive care (reduced volume tubes, near patient testing).

Before surgery: optimizing Hb and haemostasis

This process involves preassessing a patient in advance of surgery and taking steps to reduce the requirements for transfusion. ‘Preparing Patients for Surgery’ clinics can also identify medical or social reasons that may have led to an operation being cancelled and therefore increase a hospital's efficiency.

If a patient is anaemic, it is important to investigate the underlying cause. For example, iron deficiency may be due to a gastrointestinal malignancy. If a patient with iron deficiency anaemia is started on iron, the Hb can be expected to rise by about 1 g/dL per week. It is therefore important to check the blood count sufficiently far in advance of surgery to allow time for treatment to be given if required. Patients presenting for surgery with a normal Hb will require transfusion at a later stage or may even avoid blood transfusion altogether. A personal ‘goal’ is useful for the patient and their general practitioner; e.g. Hb >12 g/dL prior to elective hip replacement.

Newer formulations of intravenous iron have fewer adverse reactions than were associated with these preparations in the past. Intravenous iron is almost immediately available for red cell production. Research is being undertaken to determine whether administration of intravenous iron as late as the preoperative day can improve red cell production in response to surgical anaemia and thus decrease the use of donor blood.

It is important to consider patient factors that might cause excessive blood loss during surgery and that can be corrected in advance. Patients on aspirin or clopidogrel for secondary prevention can stop it 5–7 days before surgery (except if the patient is at high risk of suffering a myocardial infarction or has a drug-eluting coronary stent). Patients in atrial fibrillation on warfarin can discontinue the drug a few days before surgery. The newer oral anticogulants such as dagibatran and rivaroxaban are likely to rise in popularity over the next few years. For dagibatran, which has a half-life of 12–17 h, the thrombin clotting time, ecarin clotting time and TT determined by Hemoclot(R) thrombin inhibitor assay are sensitive tests to qualitatively evaluate the anticoagulant effects. The activated partial thromboplastin time (aPTT) can provide a useful qualitative assessment of anticoagulant activity but is less sensitive. The factor Xa inhibitor rivaroxaban has a shorter (7–11 h) half-life, so simple withdrawal should suffice, except for emergency surgery, when prothrombin concentrates can be used to reverse the anticoagulant effect.

If it is imperative to continue anticoagulation, such as in cases of mechanical heart valve replacement, the patient may be given intravenous heparin to cover the surgical period. It is important to take a bleeding history when the patient is seen prior to surgery. Screening tests and specific treatment may be required (see Chapter 25). Bleeding diatheses must be considered in patients with renal or liver disease. Agents such as desmopressin (DDAVP) or tranexamic acid may enhance surgical haemostasis [8] (see Chapter 37).

During surgery: reduction in blood loss

Blood loss in many operations has fallen significantly with advancing surgical and anaesthetic techniques. Use of harmonic scalpels, laparoscopy and careful surgical technique has had a huge impact on blood usage. The maintenance of normothermia ensures optimum coagulation and has also been shown to decrease blood loss.

There are several techniques specific to cardiac surgery that have been the subject of good-quality clinical trials. Protamine sulfate has an anticoagulant effect when used in excess, so reduced doses are now given following bypass surgery. The dose can be titrated using the activated clotting time, or more simply a 50% dose given. In vascular surgery, heparin is no longer reversed. Heparin-bonded bypass circuits can be used to reduce the dose of systemic heparin required. Shed mediastinal blood can be reinfused if washed (see below).

During/after surgery: when to transfuse

No blood transfusion is without risks, but equally the administration of blood may be life saving. In making the decision to transfuse, the balance of risks must be considered for each individual. Factors influencing the decision to transfuse include the Hb, the patient's life expectancy, i.e. age/prognosis (many of the adverse effects of transfusion-transmitted infection or immune modulation are delayed) and, above all, clinical judgement about the patient's ability to tolerate anaemia, including the presence of other factors such as cardiac or respiratory disease and sepsis.

Transfusion triggers

Data from patients who refuse blood on religious grounds or who live in parts of the world where blood is scarce or dangerous have helped our understanding of the effects of anaemia. In otherwise healthy patients the following transfusion triggers for stable anaemia might be considered:

·        <4 g/dL: transfuse unless fit, asymptomatic and Hb rising;

·        4–7 g/dL: transfusion usually necessary;

·        7–10 g/dL: transfusion not usually necessary; and

·        >10 g/dL: transfusion rarely required [6, 9].

A randomized trial of patients in intensive care showed that less severely ill patients (Acute Physiology and Chronic Health Evaluation II score <20) and patients under 55 years actually had a survival advantage if the Hb was maintained between 7 and 9 g/dL rather than between 10 and 12 g/dL. For patients with clinically significant cardiac disease the mortality was similar in both groups [10].

For otherwise fit patients with a previously normal Hb who are actively bleeding, the following guidelines are appropriate:

·        Blood loss <15% blood volume: give fluids; no need to transfuse.

·        Blood loss 15–30% blood volume: consider transfusion.

·        Blood loss 30–40% blood volume: transfusion usually necessary.

·        Blood loss >40% blood volume: transfusion indicated.

Note that blood volume is about 70 mL/kg in adults, so that 20% of blood volume is approximately 1 L. For patients with a short life expectancy or those with chronic anaemia and impaired red cell production, the main trigger for transfusion should be the patient's symptoms.

Transfusion targets

In addition to considering when to transfuse, a target Hb should be established for each clinical scenario using the best data available (see above and Chapter 27). It is also important to consider how many units to give. In other words, the dose of blood should depend on the estimated blood volume based on the patient's weight.

When a patient is actively bleeding, replacement of red cells should be guided by an estimate of blood loss. A guide to how many units are required to achieve the target is shown in Table 35.2. Single-unit transfusions have previously been discouraged. However, Table 35.2 shows that it might be reasonable to give one unit to a small elderly woman who is symptomatic with an Hb of 7 g/dL to bring it up to just under 9 g/dL. The transfusion of blood just because it has been made available for the patient should be avoided. If blood is not used, it can be returned to the blood bank and used for another patient. Near patient testing with a device such as the Haemocue is an essentialcomponent of modern theatre practice for incremental transfusion management. This is also gradually becoming standard for ward practice.

Table 35.2 Guide to number of units required to achieve the ‘target’ haemoglobin (Hb).

Table035-1

Techniques for providing autologous blood

ICS now seems to offer the most cost-effective method of autologous transfusion. Future issues in blood supply and demand combined with the discovery of other bloodborne diseases may change this view and result in a re-examination of PAD and ANH. Autologous blood must be clearly labelled and be distinct from donor blood. An example of an autologous blood label is shown in Figure 35.1; autologous units are more easily identified if their labels are printed a different colour to those used for allogeneic blood.

Fig 35.1 Autologous labels (provided by the UK Cell Salvage Action Group).

c35f001

Cell salvage

Principle

During surgical operations when blood loss is expected, blood can be collected, processed and then returned to the patient. This can be done either intraoperatively or postoperatively depending on the type of operation. This process can be cost effective even when small volumes of blood (i.e. more than 500 mL) are collected. The amount salvaged not only decreases the use of allogeneic blood but in many instances completely removes the need for allogeneic blood transfusion, i.e. ‘bloodless surgery’.

Intraoperative cell salvage

Intraoperative cell salvage (ICS) involves the collection and reinfusion of red cells lost during surgery [11]. This may be performed as follows:

·        Single-unit reinfusion devices (only used in fully anticoagulated patients). These are simple and cheap for low volume losses.

·        Continuous reinfusion of unprocessed blood using a dialysis technique. This may be used in conjunction with cardiac bypass but is not of proven benefit and may be associated with risk of haemolysis and high dose heparin reinfusion leading to coagulo-pathy.

·        Reinfusion of processed blood (discussed below).

There are a number of machines available that wash red cells by centrifugation and resuspend them in saline (examples are shown in Plates 35.1 to 35.4 in the plate section). Blood is aspirated from the wound site and mixed with heparin or citrate anticoagulant via dual-lumen suction before it enters the reservoir of the machine. The cycle can be either run automatically or controlled manually. In general, about 75% of red cells can be recovered for reinfusion back into the patient. The machines can deliver the equivalent of 10 units of blood per hour. Swabs laden with blood can be wrung out into a bowl of normal saline and then suctioned into the device for processing to further increase yield.

Advantages of ICS

·        There is a considerable reduction in donor blood usage in cases where blood loss is large (>1 L). Suitable operations might include open heart surgery, cystectomy and ruptured ectopic pregnancy, aortic aneurysm repair and spinal surgery (especially paediatric scoliosis correction).

·        It is available to all patients having appropriate surgery regardless of medical fitness.

·        In some situations of uncontrolled blood loss it may be life saving.

·        Unlike other techniques, ICS can be used selectively in cases where the actual, rather than the predicted, blood loss is high.

·        Blood can be collected in the reservoir and the decision to use the machine and harness can be deferred until it is clear that the blood loss is sufficient to warrant processing.

·        Cell salvage is generally accepted by Jehovah's Witnesses, provided the collected blood remains in continuity with the patient. Finally, and perhaps most importantly, the processed red cells stay by the patient's side, which eliminates almost entirely the risk of receiving the wrong blood. Identification error remains one of the biggest risks of blood transfusion.

Disadvantages/risks of ICS

Adverse events to autologous blood transfusion are now reported to the Serious Hazard of Transfusion (SHOT) scheme in the UK. In 2010 there were 15 reported events including clerical error and hypotension during reinfusion through ultrafiltration under pressure. With regard to specific concerns:

·        The reinfusion of haemolysed salvaged blood is unlikely, providing the wash process is undertaken correctly. Currently available machines operate on an automatic washing process. A sensor monitors the effluent from the wash cycle, which continues until the liquid being discarded is completely clear, suggesting removal of all free Hb, fragmented red cells and other contaminants. Quality control samples should be assayed and logged for all individual machines.

·        There have been no deaths associated with air embolism, due to improved design and greater awareness of such problems. Air embolism was only reported with very early machines, but collected blood should not be used with pressurized reinfusion devices.

·        It does not recover all the blood lost so donor blood may be required in massive haemorrhage. Platelets and coagulation factors are removed by the washing process so supplementation with allogeneic coagulation factors may be required after high volume ICS (> six cycles depending on coagulation indices) in the same way as it may be required after massive blood transfusion (see Chapter 26).

·        It requires a capital outlay and trained operators, so ICS can be used only in hospitals with sufficient numbers of suitable procedures to become cost effective. As the cost of donor blood continues to rise with the introduction of safety measures such as universal leucocyte reduction of blood and increasingly sensitive and expensive microbiology testing, cell salvage has become more cost effective.

It is important to follow agreed standard operating procedures, to document all stages of the process and maintain an audit database. Operators should be properly trained and competency assessed. In the UK this should be according to the UK Cell Salvage Action Group (UKCSAG) guidelines.

Indications for ICS

The primary indication is surgery, where expected blood loss is likely to be in excess of 500 mL. Even when blood loss is unpredictable, the collection of operative blood loss may be worthwhile. Providing this blood is anticoagulated, it can be processed and reinfused as red cells suspended in saline if sufficient volumes are collected. The processing kits are separately packaged so only the collection reservoir is wasted if small volumes are salvaged following a decision not to proceed to processing. ICS is cost-neutral providing one unit of packed red cells is reinfused. Even if a small volume of ICS blood is retransfused, raising the patient's Hb level to exceed the agreed transfusion trigger will obviate the need for donor blood.

Relative contraindications

There are a number of situations where the use of cell salvage has been discouraged. In the presence of massive haemorrhage, however, ICS may avoid hypovolaemic shock.

·        Malignant cells. Although leucocyte filters may remove the majority of cancer cells, and small numbers may not be clinically significant compared with the numbers that enter the circulation during surgery, some would advocate the use of gamma-irradiation in this setting, but this is logistically difficult to arrange. Several studies have reported large numbers of patients receiving cell salvage during cancer surgery, particularly in urology. To date there have been no reports of lung metastasis or decreased survival. The technique should be discussed on an individual basis with patients, with special arrangements for consent. The National Institute for Health and Clinical Excellence (NICE) has now approved ICS in urological malignancy in the UK, and all recipients of ICS in cancer surgery should be involved in audit or clinical trials.

·        Infection. Although the balance of risk depends on the clinical urgency for salvaged blood, antibiotics may be added to the anticoagulant solution and given parenterally to the patient to treat bacteraemia. Several trials have confirmed the value of ICS in trauma where the quality of life gain for younger fitter patients is very high.

·        Amniotic fluid in the operative field, which may cause embolism/disseminated intravascular coagulation. Studies show that circulating amniotic fluid is common during normal and Caesarean delivery. It is removed during the normal wash cycle and adverse events are rare. ICS can be life saving in complicated pregnancy such as placenta accreta. Filtering of the salvaged blood removes lamellar bodies and fetal squames and may enhance patient safety. NICE now approves ICS in obstetric practice.

·        Sickle cell disease. Cells may sickle in the machine due to low oxygen tension and therefore red cell yield would be low. This is a theoretical reason to avoid using ICS in the presence of sickle cell disease.

·        Where topical clotting agents such as fibrin glue have been used or iodine has been used to wash out the abdomen. In practice, these contaminants promote thrombin generation or haemolyse red cells. Even if these agents are collected they are washed out during the centrifugal process, but it is recommended to temporarily cease suction, irrigate the surgical area with at least 1 L of IV grade normal saline and then recommence processing.

·        There have been reports of hypotension associated with the reinfusion of ICS blood in obstetric practice, when negatively charged leucocyte depletion filters have been used. This may occur if the blood is infused under pressure. To date there are no clear data but it is important to avoid pressure reinfusion. This may limit the speed of reinfusion in cases of massive haemorrhage.

Postoperative cell salvage

Postoperative cell salvage (PCS) involves the collection of blood from surgical drains followed by reinfusion with or without processing. The blood recovered is dilute, partially haemolysed and defibrinogenated and contains high levels of cytokines unless washed. There is a clear advantage in terms of enhanced recovery following knee replacement and it may be that cellular activation and enhanced nitric oxide levels during nonwashed PCS are a positive contributory factor in boosting immunity [1]. Randomized trials comparing washed and nonwashed PCS with allogeneic blood are awaited. If the collected wound drainage blood is simply reinfused, some centres limit the quantity reinfused. Others recommend that all blood is washed and resuspended in saline. This can be done either with the apheresis machines used in the main theatre suite or with the newer and more compact processing machines that wash collected blood by the patient's bedside.

The current state of the art in postoperative drainage centres on audits to look at the precise volumes reinfused and its cost effectiveness as a blood sparing technique.

Key points

1. Autologous transfusion should be considered as part of a total quality management strategy for minimizing the risk associated with transfusion for all patients having surgery.

2. Planning and appropriate treatment in advance of or during surgery can reduce transfusion requirements.

3. An audit of ICS and PCS activity can give useful local data and inform clinicians about indicated surgical cases.

4. Before transfusing a patient always consider the strict clinical indications and how many allogeneic units are required.

5. ICS and PCS are the most effective methods of autologous transfusion.

6. The reinfused red cells are capable of carrying oxygen immediately to the tissues.

References

1. Gharehbaghian A, Haque KM, Truman C et al. Effect of autologous salvaged blood on postoperative natural killer cell precursor frequency. Lancet 2004; 363:1025–1030.

2. Goodnough LT, Brecher ME, Kanter MH & AuBuchon JP. Transfusion medicine. Part 1. N Engl J Med 1999; 340: 438–447.

3. Goodnough LT, Brecher ME, Kanter MH & AuBuchon JP. Transfusion medicine. Part 2. N Engl J Med 1999; 340: 525–533.

4. Calman KC. Cancer: science and society and the communication of risk. Br Med J 1996; 313: 799–802.

5. Consensus Statement. Autologous transfusion: 3 years on. What is new? What has happened? Transfus Med 1999; 9: 285–286.

6. American Society of Anesthesiologists Task Force on Perioperative Blood Transfusion and Adjuvant Therapies. Practice guidelines for perioperative blood transfusion and adjuvant therapies: an updated report by the American Society of Anesthesiologists Task Force on Perioperative Blood Transfusion and Adjuvant Therapies. Anesthesiology 2006; 105: 198–208.

7. British Committee for Standards in Haematology. Guidelines for policies on alternatives to allogeneic blood transfusion 1. Predeposit autologous blood donation and transfusion. Transfus Med 2007; 17: 354–365.

8. CRASH-2 Trial Collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010; 376(9734): 23–32.

9. Carless PA, Henry DA, Carson JL, Hebert PP, McClelland B & Ker K. Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion. Cochrane Database Syst Rev 2010; CD002042.

10. Hebert PC, Wells G, Blachman MA et al. A multicentre, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 1999; 340: 409–417.

11. Ashworth A & Klein A. Cell salvage as part of a blood conservation strategy in anaesthesia. Br J Anaesth 2010; 105: 401–416.

Further reading

Maniatis A, Van der Linden P & Hardy J-F (eds). Alternatives to Blood Transfusion Medicine, 2nd edn. Oxford: Wiley Blackwell; 2011.

Shander A, Javidroozi M, Ozawa S & Hare GMT. What is really dangerous: anaemia or transfusion? Br J Anaesth 2011; 107 (Suppl. 1): i41–i59.

Speiss BD, Spence RK & Shander A (eds). Perioperative Transfusion Medicine, 2nd edn. Philadelphia, PA: Lippincott Williams & Wilkins; 2006.

Thomas D, Thompson J & Ridler B (eds). A Manual for Blood Conservation. Shrewsbury: tfm Publishing Ltd; 2005.