Table of Contents Table of Contents
Previous Page  E453 E496 Next Page
Show Menu
Previous Page E453 E496 Next Page
Page Background


Vol 4 October 2017


Since the publication of the Transfusion Requirements

in Critical Care trial


in 1999, progress has been made

on the implementation of best transfusion practice on

the basis of clinical trials and clinical practice guidelines.

However, there continues to be considerable variation in

transfusion practice.


One of the factors accounting for

this variation might be disagreement about the validity

of the evidence for restrictive transfusion practice across

the many clinical settings where blood is used. A recent

Cochrane systematic review


on thresholds for red blood

cell transfusion identi ed a total of 31 randomised

controlled trials (RCTs) in 12 587 participants across

a range of clinical specialities, including surgery,

critical care, and gastrointestinal bleeding. Restrictive

transfusion strategies reduced the risk of receiving a

red blood cell transfusion by 43% (risk ratio [RR] 0·57,

95% CI 0·49–0·65). Overall, restrictive transfusion

strategies did not increase or decrease the risk of 30-day

mortality when compared with liberal transfusion

strategies (RR 0·97, 95% CI 0·81–1·16). Additionally,

restrictive transfusion strategies did not a ect any of

the other assessed outcomes including cardiac events,

myocardial infarction, stroke, thromboembolism, or

infection. These ndings provide good evidence that

allogeneic blood transfusions are of no bene t and can

be avoided in most haemodynamically stable patients

with haemoglobin concentrations of more than

70–80 g/L. Consideration of the presence or absence

of comorbidities such as pre-existing cardiovascular

disease has been recommended in decision making

about the need for transfusion, speci cally at a

haemoglobin threshold of 80 g/L versus 70 g/L.


However, available data are insu cient to inform

the safety of restrictive transfusion policies in certain

clinical subgroups, including acute coronary syndrome,

myocardial infarction, traumatic brain injury, and acute

neurological disorders such as stroke.

Another risk factor that could a ect decision

making on the optimal red blood cell transfusion

threshold is patient age. In a systematic review and

meta-analysis published in

The Lancet Haematology


Geo Simon and colleagues


analysed the clinical

outcomes of patients aged 65 years or older in RCTs of

liberal versus restrictive transfusion strategies. Nine

trials consisting of 5780 participants were identi ed:

ve in an orthopaedic setting, three in cardiology, and

one in oncology. The 30-day mortality (RR 1·36, 95%

CI 1·05–1·74; p=0·017) and 90-day mortality outcomes

(1·45, 1·05–1·98; p=0·022) were lower in older patients

who followed a liberal transfusion strategy than in

those who followed a restrictive strategy. The authors

concluded that their data challenge the application of

current guidelines for restrictive transfusion strategies

in elderly patients. The care of elderly patients receiving

more liberal transfusions might increase the demand

for blood as transfusion requirements increase

progressively, especially after the age of 65 years.


Before accepting these conclusions and putting

them into practice, the limitations of the meta-analysis

should be examined. Only three of the nine included

RCTs, with a total of 590 patients, were restricted

to patients aged 65 years and older. The other trials

included patients starting at 16 years or 18 years of

age, although the median age in all studies was at

least 64 years. However, we found that the authors

did not include 11 trials ( ve orthopaedic, two cardiac,

two critical care, one vascular, and one gastrointestinal

bleeding: 2943 participants) from the Cochrane review


that all had a mean or median participant age of at

least 65 years. If data from these trials are included in

the analysis of 30-day mortality, there is no longer a

signi cant di erence in this outcome (appendix). This

nding puts in to question all of the results from Simon

and colleagues’ review


because at least 50% of the data

from eligible participants are missing.

Nevertheless, the authors should be congratulated on

raising concern about the general adoption of restrictive

red blood cell transfusion for elderly patients based on

the results of RCTs, which have included patients with

broad age ranges. Several questions about optimal

blood transfusion thresholds remain to be answered,

and not just in relation to the clinical subgroups

identi ed in the Cochrane review.


One way of achieving

answers would be through new trials designed to

develop the evidence base, such as restrictive versus

liberal red blood cell transfusion in elderly patients.

Another approach would be to compare data for

patients younger or older than 65 years, with or without

known cardiovascular disease, from published RCTs of

restrictive versus liberal transfusion strategies. In the

Blood transfusion strategies in elderly patients



September 11, 2017 S2352-3026(17)30173-4


Online/Correspondence S2352-3026(17)30172-2



page e465



for appendix

Life InView/SciencePhotoLibrary