Restrictive vs. Liberal Red Blood Cell Transfusion Thresholds: A Comprehensive Evidence Review Scope and Framing This review covers the full evidence landscape for restrictive versus liberal RBC transfusion strategies across clinical populations — critically ill, cardiac surgery, acute myocardial infarction, GI bleeding, orthopedic surgery, and emerging domains (neurocritical care, hematologic malignancies). The…
THE EVIDENCE
What each source contributes, and how strong it is
Study design first — a cohort and a randomised trial do not carry the same weight.
Design
Year
Journal
What it found
Meta-analysis
2025
NEJM evidence
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial
Randomised trial
2011
The New England journal of medicin
In FOCUS, a 10 g/dL threshold after hip fracture surgery in patients at
Randomised trial
1999
The New England journal of medicin
In TRICC, transfusing at a haemoglobin below 7 g/dL was at least as effective as
Cohort
2025
Transfusion
Transfusing patients whose oxygen extraction ratio exceeded 0.30 improved that ratio
Review
2026
Intensive care medicine
Haemoglobin alone does not reflect tissue oxygenation
Guideline
2025
Annals of internal medicine
For hospitalised patients with acute myocardial infarction the AABB guideline
6 resolved citations behind this deck; every point above traces to one of them.
WHERE THE GUIDANCE SITS
What the guidelines and pooled evidence say
2 findings, each on the slide that follows.
NEJM evidence 2025
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial infarction gave more cardiac death at 30 days (5.5% versus 3.7%) and…
Annals of internal medicine 2025
For hospitalised patients with acute myocardial infarction the AABB guideline
For hospitalised patients with acute myocardial infarction the AABB guideline suggests a liberal strategy below 10 g/dL, noting a restrictive 7-8…
WHERE THE GUIDANCE SITS
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial
Meta-analysis · NEJM evidence
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial infarction gave more cardiac death at 30 days (5.5% versus 3.7%) and higher all-cause mortality at 6 months.
Restrictive versus Liberal Transfusion in Myocardial Infarction — A Patient-Level Meta-Analysis, NEJM evidence 2025 · PMID 39714935
THE NUMBERS
In myocardial infarction, restrictive transfusion looked worse
4,311 patients pooled at individual level, including MINT.
Every point favours the liberal strategy here — the opposite of the general critical-care teaching.
Restrictive versus Liberal Transfusion in Myocardial Infarction — A Patient-Level Meta-Analysis, NEJM evidence 2025 · PMID 39714935
WHERE THE GUIDANCE SITS
For hospitalised patients with acute myocardial infarction the AABB guideline
Guideline · Annals of internal medicine
For hospitalised patients with acute myocardial infarction the AABB guideline suggests a liberal strategy below 10 g/dL, noting a restrictive 7-8 g/dL threshold may increase mortality in this group.
Red Cell Transfusion in Acute Myocardial Infarction: AABB International Clinical Practice Guidelines, Annals of internal medicine 2025 · PMID 40825204
WHAT THE TRIALS FOUND
Where randomised evidence moved the question
2 findings, each on the slide that follows.
The New England journal of medicine 2011
In FOCUS, a 10 g/dL threshold after hip fracture surgery in patients at
In FOCUS, a 10 g/dL threshold after hip fracture surgery in patients at cardiovascular risk gave no better survival or independent walking at 60 days…
The New England journal of medicine 1999
In TRICC, transfusing at a haemoglobin below 7 g/dL was at least as effective as
In TRICC, transfusing at a haemoglobin below 7 g/dL was at least as effective as below 10 g/dL in critically ill patients, with lower in-hospital…
WHAT THE TRIALS FOUND
In FOCUS, a 10 g/dL threshold after hip fracture surgery in patients at
Randomised trial · The New England journal of medicin
In FOCUS, a 10 g/dL threshold after hip fracture surgery in patients at cardiovascular risk gave no better survival or independent walking at 60 days than transfusing at 8 g/dL or for symptoms.
Liberal or Restrictive Transfusion in High-Risk Patients after Hip Surgery, The New England journal of medicine 2011 · PMID 22168590
THE NUMBERS
FOCUS: after hip fracture, the higher threshold gained nothing
2,016 patients at cardiovascular risk.
No benefit from transfusing at the higher threshold in a high-risk orthopaedic population.
Liberal or Restrictive Transfusion in High-Risk Patients after Hip Surgery, The New England journal of medicine 2011 · PMID 22168590
WHAT THE TRIALS FOUND
In TRICC, transfusing at a haemoglobin below 7 g/dL was at least as effective as
Randomised trial · The New England journal of medicin
In TRICC, transfusing at a haemoglobin below 7 g/dL was at least as effective as below 10 g/dL in critically ill patients, with lower in-hospital mortality and the possible exception of acute myocardial infarction and unstable angina.
A Multicenter, Randomized, Controlled Clinical Trial of Transfusion Requirements in Critical Care, The New England journal of medicine 1999 · PMID 9971864
THE NUMBERS
TRICC: restrictive was at least as good in general critical illness
838 critically ill patients.
The foundation of restrictive practice — with acute myocardial infarction flagged as the exception even in 1999.
A Multicenter, Randomized, Controlled Clinical Trial of Transfusion Requirements in Critical Care, The New England journal of medicine 1999 · PMID 9971864
IN PRACTICE
What the cohorts and reviews add
2 findings, each on the slide that follows.
Transfusion 2025
Transfusing patients whose oxygen extraction ratio exceeded 0.30 improved that ratio
Transfusing patients whose oxygen extraction ratio exceeded 0.30 improved that ratio and central venous oxygen saturation
Intensive care medicine 2026
Haemoglobin alone does not reflect tissue oxygenation
Haemoglobin alone does not reflect tissue oxygenation
IN PRACTICE
Transfusing patients whose oxygen extraction ratio exceeded 0.30 improved that ratio
Cohort · Transfusion
Transfusing patients whose oxygen extraction ratio exceeded 0.30 improved that ratio and central venous oxygen saturation, while those at or below 0.30 gained no physiological benefit.
Red Blood Cell Transfusion in the Intensive Care Unit, Transfusion 2025 · PMID 40134137
IN PRACTICE
Haemoglobin alone does not reflect tissue oxygenation
Review · Intensive care medicine
Haemoglobin alone does not reflect tissue oxygenation, and a single threshold across an ICU stay prevents neither unnecessary nor insufficient transfusion; central venous oxygen saturation and oxygen extraction ratio are bedside alternatives under study.
Physiological Determinants and the Red Blood Cells Transfusion Decision-Making Process in Non-Bleeding Critically Ill Patients: A Comprehensive Narrative Review, Intensive care medicine 2026 · PMID 41586889
IN THE ROOM
What this changes about the next case
Colour is the strength of the evidence behind each step, not the urgency.
1
1
For hospitalised patients with acute myocardial infarction the AABB guideline suggests a liberal strategy below 10 g/dL, noting a restrictive 7-8 g/dL threshold may increase mortality in this group.
2
2
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial infarction gave more cardiac death at 30 days (5.5% versus 3.7%) and higher all-cause mortality at 6 months.
3
3
In FOCUS, a 10 g/dL threshold after hip fracture surgery in patients at cardiovascular risk gave no better survival or independent walking at 60 days than transfusing at 8 g/dL or for symptoms.
4
4
In TRICC, transfusing at a haemoglobin below 7 g/dL was at least as effective as below 10 g/dL in critically ill patients, with lower in-hospital mortality and the possible exception of acute myocardial infarction and unstable angina.
5
5
Transfusing patients whose oxygen extraction ratio exceeded 0.30 improved that ratio and central venous oxygen saturation, while those at or below 0.30 gained no physiological benefit.
The oral-boards stem on the next slide puts these into one scenario.
KEY TAKEAWAYS
What to carry into the next case
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial
NEJM evidence 2025
In FOCUS, a 10 g/dL threshold after hip fracture surgery in patients at
The New England journal of medicine 2011
In TRICC, transfusing at a haemoglobin below 7 g/dL was at least as effective as
The New England journal of medicine 1999
Transfusing patients whose oxygen extraction ratio exceeded 0.30 improved that ratio
Transfusion 2025
Pooling 4,311 patients including MINT, restrictive transfusion in myocardial
Questions I'll ask you in the room
How does your plan change if the patient is not optimised?
What is the physiology behind what we just did?
Talk me through the trade-off you made there.
What would make you abandon this plan and do something else?
Oral boards stem
A 70-year-old with known coronary disease is midway through a long urological case. The haemoglobin returns at 7.6 g/dL, he is haemodynamically stable, and the resident asks whether to transfuse.
Board questions
1. For hospitalised patients with acute myocardial infarction, state the transfusion strategy and haemoglobin threshold suggested by the AABB international clinical practice guideline, and what that guideline notes about a restrictive 7-8 g/dL threshold in this group.
Show answer
For hospitalised patients with acute myocardial infarction the AABB guideline suggests a liberal strategy below 10 g/dL, noting that a restrictive 7-8 g/dL threshold may increase mortality in this group.
Acute myocardial infarction is the named exception in the AABB guideline: a liberal strategy below 10 g/dL, because the restrictive 7-8 g/dL threshold may increase mortality here.
Red Cell Transfusion in Acute Myocardial Infarction: AABB International Clinical Practice Guidelines, Annals of internal medicine 2025 · PMID 40825204
2. A patient-level meta-analysis pooled 4,311 patients, including those from the MINT trial, comparing restrictive with liberal transfusion in myocardial infarction. State what restrictive transfusion produced for cardiac death at 30 days, giving both rates, and for all-cause mortality at 6 months.
Show answer
Restrictive transfusion in myocardial infarction gave more cardiac death at 30 days, 5.5% versus 3.7%, and higher all-cause mortality at 6 months.
Myocardial infarction is the setting where the restrictive habit turns against you: more cardiac death at 30 days and higher all-cause mortality at 6 months in this pooled patient-level analysis.
Restrictive versus Liberal Transfusion in Myocardial Infarction — A Patient-Level Meta-Analysis, NEJM evidence 2025 · PMID 39714935
3. In the TRICC trial, critically ill patients were transfused for a haemoglobin below 7 g/dL or below 10 g/dL. State the trial's conclusion about the lower threshold, what was seen for in-hospital mortality, and which patient groups were flagged as a possible exception.
Show answer
TRICC concluded that transfusing at a haemoglobin below 7 g/dL was at least as effective as below 10 g/dL in critically ill patients, with lower in-hospital mortality, with the possible exception of acute myocardial infarction and unstable angina.
TRICC is where the restrictive threshold came from, and note that it carved out the possible exception of acute myocardial infarction and unstable angina from the very beginning.
A Multicenter, Randomized, Controlled Clinical Trial of Transfusion Requirements in Critical Care, The New England journal of medicine 1999 · PMID 9971864
4. A resident proposes adopting a single haemoglobin trigger for every non-bleeding ICU patient for the whole admission. According to a narrative review of the physiological determinants of transfusion decision-making in non-bleeding critically ill patients, state the two limitations of that approach and the current status of the proposed bedside alternatives.
Show answer
Haemoglobin alone does not reflect tissue oxygenation, and a single threshold across an ICU stay prevents neither unnecessary nor insufficient transfusion; central venous oxygen saturation and oxygen extraction ratio are bedside alternatives that are still under study rather than established replacements.
A haemoglobin number is not a measure of tissue oxygenation, and one fixed trigger for a whole ICU stay will both over- and under-transfuse, which is why physiologic markers like ScvO2 and oxygen extraction ratio are being studied.
Physiological Determinants and the Red Blood Cells Transfusion Decision-Making Process in Non-Bleeding Critically Ill Patients: A Comprehensive Narrative Review, Intensive care medicine 2026 · PMID 41586889
The bottom linePooling 4,311 patients including MINT, restrictive transfusion in myocardial
Sources
[1] Restrictive versus Liberal Transfusion in Myocardial Infarction — A Patient-Level Meta-Analysis, NEJM evidence 2025 · PMID 39714935 open
[2] Liberal or Restrictive Transfusion in High-Risk Patients after Hip Surgery, The New England journal of medicine 2011 · PMID 22168590 open
[3] A Multicenter, Randomized, Controlled Clinical Trial of Transfusion Requirements in Critical Care, The New England journal of medicine 1999 · PMID 9971864 open
[4] Red Blood Cell Transfusion in the Intensive Care Unit, Transfusion 2025 · PMID 40134137 open
[5] Physiological Determinants and the Red Blood Cells Transfusion Decision-Making Process in Non-Bleeding Critically Ill Patients: A Comprehensive Narrative Review, Intensive care medicine 2026 · PMID 41586889 open
[6] Red Cell Transfusion in Acute Myocardial Infarction: AABB International Clinical Practice Guidelines, Annals of internal medicine 2025 · PMID 40825204 open