Red cell transfusions for treating anaemia in the absence of bleeding

BMJ 2015; 350 doi: https://doi.org/10.1136/bmj.h1463 (Published 24 March 2015) Cite this as: BMJ 2015;350:h1463
  1. Timothy S Walsh, professor of critical care
  1. 1Department of Anaesthesia, Critical Care and Pain Medicine, Royal Infirmary of Edinburgh, Edinburgh University, Edinburgh EH16 4SA, UK
  1. twalsh{at}staffmail.ed.ac.uk

Use a restrictive strategy in most patient groups

Transfusion of donated (allogeneic) red blood cells is life saving for patients with major bleeding, such as after major trauma or childbirth. However, probably less than 15% of the 85 million units of blood transfused globally each year are used to treat haemorrhagic shock; most are given to patients experiencing blood loss during surgery, or with chronic medical conditions or haematological malignancies.1 In these situations the circulating blood volume is either normal or can be restored with other fluids such as crystalloid solutions. For these patients it is uncertain when the benefits of supplementing oxygen carrying capacity by transfusing allogeneic blood outweigh the risks. In a linked updated meta-analysis Holst and colleagues (doi:10.1136/bmj.h1354) explore this issue by comparing more liberal with more restrictive transfusion practices and suggest that in many clinical situations a restrictive transfusion strategy does no harm and reduces patients’ exposure to potentially hazardous blood products.2

Red blood cells transport oxygen to cells. Non-anaemic adults transport about 1 litre of oxygen from the lungs to tissues every minute but use only 0.25 litres, creating a wide “safety margin.” When tissues acutely need more oxygen (for example during exercise), the cardiac output and the proportion of oxygen extracted from blood increase to maintain supply. The “anaerobic threshold,” beyond which the lactate concentration in blood increases, is rarely reached except for short self limiting periods. In anaemic patients without hypovolaemia, a reduced blood viscosity and activation of the sympathetic nervous system increase cardiac output and maintain oxygen supply. Few patients with a haemoglobin concentration greater than 70 g/L have an increased plasma lactate level, and when lactic acidosis is present it is often associated with other conditions that interfere with cellular oxygen use, such as sepsis or systemic inflammation.3 Controlled experiments in healthy humans show that anaerobic metabolism is difficult to detect even at a haemoglobin concentration of 50 g/L.4 Although anaemia causes fatigue and other non-life threatening symptoms, treating the underlying cause rather than transfusing blood is often effective over several weeks. Even in critically ill anaemic patients it is difficult to demonstrate a physiological benefit from transfusions at the “whole body” level.5

Many clinicians worry that some organs, especially the heart, are more susceptible to anaemia than others. Heart muscle extracts about 60% of oxygen from the coronary circulation, a higher proportion than other major organs, and must increase coronary blood flow to meet increased oxygen requirements. If coronary flow is compromised by acute or chronic coronary disease, it is physiologically plausible that anaemia could increase ischaemia and that transfusing at a higher (uncertain) haemoglobin concentration may reverse or prevent this ischaemia. Observational studies report wide variation in clinical practice, but many indicate possible harm from blood transfusion even in this population, consistent with most observational transfusion studies in other patients.6

Blood transfusions are safer now than they have ever been because of ever increasing (and expensive) screening, testing, and processing procedures and national programmes to reduce and monitor administration errors and transfusion related adverse events. Despite these advances adverse effects from transfusion remain biologically plausible: donor leucocytes could modify recipient immunity and increase the risk of infection, and changes during storage could decrease the oxygen carrying capacity of red cells or even activate proinflammatory or prothrombotic pathways.7 Standards for red cell production focus on excluding pathogens and ensuring stored red cells are viable after transfusion, neither of which relate directly to their efficacy as oxygen transporters. It is ironic that the effectiveness and safety of a treatment used worldwide in all age groups and almost all areas of healthcare has never been formally subjected to the same scrutiny as have novel drugs or technologies. There are also virtually no cost effectiveness data to guide transfusion policies, despite the enormous global annual expenditure on production. History partly explains why this has occurred, but we have a responsibility to donors, recipients, and tax payers to improve our understanding of when anaemic patients benefit from transfusion, and when they may be harmed.

Holst and colleagues report an updated systematic review and meta-analysis of randomised trials comparing more liberal with more restrictive transfusion practices.2 New evidence includes large trials in two relatively homogeneous populations—namely, patients with upper gastrointestinal bleeding8 and critically ill patients with septic shock.9 Both conditions have a high mortality, and impaired oxygen delivery is a potential contributory factor. Holst and colleagues found no evidence of excess mortality, adverse events, or myocardial infarctions when transfusions were restricted to a haemoglobin threshold of 70-80 g/L, an effect observed overall, in each of only five trials that recruited more than 500 participants, and in pooled analyses confined to the highest quality trials. Restricting transfusions reduced infections, consistent with another recent meta-analysis.10 The number of participants was fewer than 10 000 across populations with wide ranging types of illness and severity and undergoing procedures with a widely varying risk of death. This heterogeneity supports the authors’ conclusion that more information is needed, especially in relation to myocardial events and mortality.

So which populations should be the focus of future research? The data clearly show us that patients with minimal comorbidity, low illness severity, or undergoing low risk procedures do not benefit from transfusions when their haemoglobin concentration is more than 70 g/L. In my view these patients should be excluded from future trials and the focus should be on implementing best practice.11 Uncertainty remains for patients with acute coronary syndromes, including those with acute coronary vessel disease (type 1 myocardial infarction), but also those with ischaemia from an imbalance in coronary supply-demand (type 2 myocardial infarction). Type 2 myocardial infarction could occur, for example, in patients with stable coronary disease who develop another illness such as sepsis. In inconclusive underpowered subgroup analyses (both post hoc12 and prespecified9) of the two largest critical care trials, more deaths occurred in the restrictive than liberal transfusion groups in patients with pre-existing ischaemic heart disease. There are no large trials comparing different transfusion strategies in adults with acute (type 1) myocardial infarction.

As the population ages and the prevalence of comorbidity increases, we still need more high quality research to ensure cost effective use of a precious, expensive, and potentially harmful resource to ensure blood transfusions are targeted at those patients most likely to benefit.


Cite this as: BMJ 2015;350:h1463


  • Research, doi:10.1136/bmj.h1354
  • Competing interests: I have read and understood the BMJ policy on declaration of interests and declare the following: none.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract

Log in

Log in through your institution


* For online subscription