Secondary Logo

Journal Logo


How can we explain the gap between randomised studies and ‘real life’ practice in postoperative transfusion triggers? Do we need to change recommended thresholds for transfusion?

Rosencher, Nadia; Ozier, Yves; Souied, François; Lienhart, André; Samama, Charles-Marc

Author Information
European Journal of Anaesthesiology: October 2012 - Volume 29 - Issue 10 - p 460-461
doi: 10.1097/EJA.0b013e328356073c

At the present time, improvements in blood safety, particularly in terms of transfusion-transmitted viral infections, have tended to trivialise the risks of transfusion, suggesting that the risk of no transfusion or a delay in transfusion appears to be much higher than the risk of transfusion itself. Indeed, in a French survey performed on anaesthesia-related mortality,1 the estimated number of anaesthetic procedures was 7756 121 per year, with a total of 419 deaths totally or partly related to anaesthesia. Several common causes of death were identified, but consequences of haemorrhage and anaemia played a disturbing role. It was estimated from this survey that nearly 100 deaths occur perioperatively in France every year as the result of inadequate blood management. Surprisingly, more deaths occurred which were related partly to delayed or absent blood transfusion, and only a small proportion of complications occurred after an episode of transfusion, emphasising the safety of blood transfusion in contemporary practice. Moreover, 58% of the reported deaths occurred more than 24 h after surgery. In the presence of massive blood loss, anaemia can lead to myocardial infarction (MI) and cardiovascular deaths, which are still the most frequent causes of death after noncardiac surgery, as demonstrated in the huge Norwegian Register of more than 60 000 elective orthopaedic operations.2,3 Patients undergoing noncardiac surgery are at risk of adverse perioperative complications and impaired long-term outcome.

Should we not, therefore, modify the recommended transfusion thresholds, especially when these theoretical limits are not always followed?

In terms of mortality related to transfusion triggers, cohort, observational and randomised studies give conflicting results. In the first adequately powered randomised controlled trial (RCT) in an ICU,4 Hebert compared two groups of patients: a restrictive threshold to maintain haemoglobin (Hb) concentration between 7–9 g dl−1 and a liberal transfusion threshold to maintain Hb between 10–12 g dl−1. The incidence of death from all causes in the 30 days after admission to the ICU was 18.7% in the restrictive-strategy group and 23.3% in the liberal-strategy group. However, no information was provided about the subgroup of coronary patients who were excluded from the study.

Can we compare ICU patients and postoperative patients who require active rehabilitation? Early rehabilitation before discharge was responsible for a 50% reduction in mortality at 6 months in patients with hip fracture; consequently, active rehabilitation is needed.5 In patients with hip fracture, two RCTs have shown different results. The study by Foss et al.6 included 120 patients and showed a significantly higher mortality in the restrictive group (8–10 g dl−1) compared with the liberal group (10 g dl−1), but no difference in active rehabilitation, walking fatigue or dizziness. The study by Carson et al.7 of 2000 patients after hip fracture showed no difference in terms of mortality or morbidity between patients in the restrictive group (not transfused if Hb ≥8 g dl−1) and the liberal group (transfusion to reach Hb = 10 g dl−1). This important study calls for some comments. First, Hb in the restrictive group remained around 9 g dl−1 (not 8 g dl−1) on average throughout the study, and in the liberal group, the mean Hb was higher than 10 g dl−1 throughout the study. Second, there was a nonsignificant trend towards a higher rate of MI in the restrictive group. Third, in this group, there was a significantly greater number of patients who required transfusion as a result of clinical symptoms (such as tachycardia). It should be noted that all patients in the restrictive group who had a sign of intolerance to anaemia were transfused, thus reducing the mortality rate in this group.

Therefore, the debate should now focus on the following question: should we wait for clinical signs of intolerance to anaemia before starting transfusion? The answer is controversial for many reasons. According to the Perioperative Ischaemic Evaluation study, 5% of patients undergoing noncardiac surgery had a perioperative MI within 30 days of randomisation.8 Most (74.1%) of these MIs occurred within 48 h of surgery. Moreover, 65.3% of patients did not experience any ischaemic symptoms. The 30-day mortality rate was 11.6% of patients who had a perioperative MI (48 of 415 patients) but only 2.2% (178 out of 7936 patients) among those who did not (P < 0.001). It is important to emphasise that mortality rates were high and quite similar between those with (9.7%) and without ischaemic symptoms (12.5%).8

Is it reasonable to imagine that a physician should be available at any time of the day in every ward to look after every patient in order to diagnose any episode of anaemia-related tachycardia? Obviously, this only happens in RCTs. In real life, what is the mean time period to obtain one unit of red blood cells (RBCs) and to transfuse it? If the blood sample is collected at 08.00 h, results are most often seen at 11.00 h and transfusion may be completed between 14.00 and 16.00 h. Therefore, if the patient is still bleeding, this situation can lead to an ischaemic episode. Only knowledge and analysis of logistical problems involved in obtaining a unit of RBC can improve the process. Knowing the kinetics of bleeding for each procedure will help anticipate transfusion needs in line with the recommended limits. All records show that MI is the leading cause of postoperative mortality. Postoperative independent predictors are renal failure and bleeding.8 Bleeding causes severe anaemia, itself responsible for MI, if transfusion is delayed too long.

The recommendations are still valid, but we must now move on towards tailoring RBC administration to suit individual needs by adapting to ward routines (delay for obtaining a unit of RBC from the blood bank), to logistical problems of obtaining blood in a timely fashion and to the kinetics of bleeding for each procedure. We must still remember that transfusion is associated with potential complications and its use must be undertaken with care. It should not be limited to knowledge of the medical history; we must also put into perspective the expected kinetics of blood loss and its possible unexpected volume. We should not manage anaemia in the recovery room and at postoperative day 5 in the same way. Sometimes, according to the kinetics of bleeding, the transfusion trigger should be different in the recovery room and in the ward. Today, point-of-care devices for Hb measurement must find their way into surgical wards, in addition to clinical monitoring. Finally, it is the responsibility of the medical team to implement measures to control Hb and transfusion requirements in a timely manner.


Assistance with the Editorial: none declared.

Sources of funding: none declared.

Conflicts of interest: none declared.

CMS is an associate editor of the European Journal of Anaesthesiology.

Comment from the Editor: This article was checked and accepted by the Editors, but was not sent for external peer-review.


1. Lienhart A, Auroy Y, Pequignot F, et al. Survey of anesthesia-related mortality in France. Anesthesiology 2006; 105:1087–1097.
2. Lie SA, Engesaeter LB, Havelin LI, et al. Early postoperative mortality after 67 548 total hip replacements: causes of death and thromboprophylaxis in 68 hospitals in Norway from 1987 to 1999. Acta Orthop Scand 2002; 73:392–399.
3. Mantilla CB, Horlocker TT, Schroeder DR, et al. Frequency of myocardial infarction, pulmonary embolism, deep venous thrombosis, and death following primary hip or knee arthroplasty. Anesthesiology 2002; 96:1140–1146.
4. Hebert PC, Wells G, Blajchman MA, et al. A multicenter, 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.
5. Rosencher N, Vielpeau C, Emmerich J, et al. Venous thromboembolism and mortality after hip fracture surgery: the ESCORTE study. J Thromb Haemost 2005; 3:2006–2014.
6. Foss NB, Kristensen MT, Jensen PS, et al. The effects of liberal versus restrictive transfusion thresholds on ambulation after hip fracture surgery. Transfusion 2009; 49:227–234.
7. Carson J, Terrin MD, Noveck H, et al. Liberal or restrictive transfusion in high-risk patients after hip surgery. New Engl J Med 2011; 365:2453–2462.
8. Devereaux PJ, Xavier D, Pogue J, et al. Characteristics and short-term prognosis of perioperative myocardial infarction in patients undergoing noncardiac surgery: a cohort study. Ann Intern Med 2011; 154:523–528.
© 2012 European Society of Anaesthesiology