Skip to content
View commentaries on this research

This is a plain English summary of an original research article

For people with existing heart disease or stroke, a strategy of transfusing blood if anaemia is less severe, haemoglobin above 80g/L, led to fewer coronary syndromes. The strategy was compared to an alternative where people received a blood transfusion only if they had more severe anaemia, mostly haemoglobin levels less than 80g/L.

This systematic review of randomised trials looked at the rates or coronary syndromes, heart attacks or episodes of unstable angina, in adults with past cardiovascular conditions - hospitalised for any reason other than heart surgery. The rates of complications for both strategies was low; 4.6 episodes of acute coronary syndrome for every 100 patients transfused when the haemoglobin level was 70 to 80g/L and 2.7 coronary syndromes for every 100 patients transfused in the typical range 90 to 100g/L, but the difference is important. Deaths in the two groups were no different.

Current guidance recommends that people admitted with suspected heart attacks should be treated at levels signifying less severe anaemia (above 80g/L), known as a liberal strategy. These new findings suggest that people with non-acute cardiovascular disease may also benefit from being managed the same way.

Why was this study needed?

About seven million people in the UK have cardiovascular disease, and it is a common comorbidity among patients admitted to hospital. Anaemia, low levels of red blood cells or haemoglobin, is associated with poorer patient outcomes, but the level of anaemia at which transfusion is indicated depends on the individual and their disease. It is thought that haemoglobin levels of around 70 g/L are safe for most patient groups.

However, it is not clear what the best transfusion strategy is for people with existing cardiovascular disease, who may be at greater risk from the reduced oxygen supply that comes with anaemia.

No systematic reviews have specifically compared outcomes for patients with cardiovascular disease who are in hospital for reasons other than heart surgery. Guidelines, including those from NICE, acknowledge the lack of quality evidence in this area.

What did this study do?

This was a systematic review and meta-analysis of 11 randomised controlled trials that compared ‘restrictive strategies’, cut-off values ranged from 70 to 97g/L, with ‘liberal strategies’, cut off values from 90 to 113g/L, for adults with anaemia and existing cardiovascular disease.

The setting and populations included in the trials varied: four were in critical care; three were in orthopaedics, one in vascular surgery, one in people with upper gastrointestinal bleeding, and two in people with acute coronary syndrome (heart attack or unstable angina).

Six of the trials were judged to be at high risk of bias due to the lack of blinding of participants and staff, though this would have been hard to do. The pooling of overlapping cut-off values made comparison of the restrictive and liberal groups challenging. Also definitions of cardiovascular disease differed between trials. However, heterogeneity across trials was found to be low, increasing our confidence that the meta-analysis was valid.

What did it find?

  • There was no difference in 30 day mortality between groups. 9.5% of people in the restrictive transfusion group died compared with 8% in the liberal transfusion group (risk ratio [RR] 1.15, 95% confidence interval [CI] 0.88 to 1.50; 11 trials including 3033 adults).
  • Patients managed with a restrictive transfusion threshold had a greater risk of acute coronary syndrome (heart attack or unstable angina), which occurred in 4.5% of the restrictive transfusion group compared to 2.5% in the liberal transfusion group (RR 1.78, 95% CI 1.18 to 2.70; 9 trials including 2609 adults).
  • A subgroup analysis excluding the two trials of people hospitalised with acute coronary syndrome had no significant impact on the findings.
  • Six trials found no difference in length of hospital stay between the groups. Other outcomes were too rare to be pooled in meta-analysis.

What does current guidance say on this issue?

NICE guidance published in 2015 recommends using a restrictive transfusion threshold (70g/L) unless patients have acute coronary syndrome, major bleeding or chronic anaemia. For people with acute coronary syndrome the higher haemoglobin level of 80g/L is advised. There is no specific recommendation for patients with cardiovascular disease who are not having an acute heart event, and highlights the need for this research.

The UK National Blood Transfusion Committee recommends the use of locally agreed triggers for transfusion based on national guidelines. They also recommend the development of systems that empower laboratory staff to question requests that do not conform to the agreed triggers, and regular local audits of transfusion requests.

What are the implications?

NICE already recommend a liberal transfusion threshold of 80 g/L for patients with acute coronary syndrome. These findings suggest that people with existing cardiovascular disease who are not experiencing an acute heart event may also benefit from being treated if they develop less severe anaemia (haemoglobin level 80 to 100g/L) a liberal strategy.

A recent Signal covered a trial indicating that restrictive transfusion strategy may not be safe for people undergoing heart surgery and that the higher haemoglobin threshold of 80mg/L might be better.

The limitations to the review are unlikely to change the main conclusions that support the more liberal threshold, but the exact threshold may change in further research. For example because some trials included people with an acute heart attack, some with heart disease, and some with cardiovascular risk factors only (e.g. high blood pressure) or other vascular disease such as stroke, it becomes hard to define for whom the threshold applies. In many of the trials, the people assessing outcomes were not blinded to the treatment groups, and the restrictive and liberal transfusion thresholds varied and overlapped between trials.

However, the direction of effect across trials consistently indicated that a liberal transfusion threshold reduced risk of new acute heart events, such as a heart attack or unstable angina. So this is probably the best randomised evidence that there is likely to be in this area.

Citation and Funding

Docherty AB, O'Donnell R, Brunskill S, et al. Effect of restrictive versus liberal transfusion strategies on outcomes in patients with cardiovascular disease in a non-cardiac surgery setting: systematic review and meta-analysis. BMJ. 2016;352:i1351.

No funding information was provided for this study.


National Clinical Guideline Centre. Transfusion. Blood transfusion: NICE guideline NG24. London: Royal College of Physicians; 2015.

NHS Blood and Transplant. About blood. London: National Health Service; 2015.

NICE. Blood transfusion. NG24. London: National Institute for Health and Care Excellence; 2015.

Royal College of Surgeons. Surgery and the NHS in numbers. London: The Royal College of Surgeons of England; 2016.

Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre

  • Share via:
  • Print article


Acute coronary syndrome describes a series of conditions where the underlying cause is a significant blockage in the coronary arteries. There are three main types of acute coronary syndrome, as described below.

1) Heart attack (ST segment elevation): The most serious type of heart attack, where there is a long interruption to the blood supply. This is caused by a total blockage of the coronary artery, which can cause extensive damage to a large area of the heart.

2) Heart attack (non-ST segment elevation): Less serious than above because the supply of blood to the heart is only partially, rather than completely, blocked. As a result, a smaller section of the heart is damaged.

3) Unstable angina: The least serious type of acute coronary syndrome although, like a heart attack, it is still regarded as a medical emergency. In unstable angina, the blood supply to the heart is still seriously restricted, but there is no permanent damage, so the heart muscle is preserved.


Back to top