No separate transfusion threshold for the critically ill patient with infarction

Design

Prespecified subgroup analysis of a two-arm randomised controlled trial across 144 sites in six countries

Population

3,504 adults with acute myocardial infarction and anaemia; 1,679 (47.9%) in intensive care at randomisation

Primary outcome

Death, and the composite of death or myocardial infarction, at 30 days

Effect

Restrictive vs liberal: death risk ratio 1.09 (95% CI 0.77-1.54) in the non-critically ill and 1.24 (0.95-1.61) in the critically ill; death or reinfarction 1.09 (0.85-1.38) and 1.21 (0.99-1.47); interaction p = 0.55 and 0.52

Restrictive transfusion is the default in critical care, but the old TRICC data raised the possibility that it harms patients with ischaemic heart disease — leaving intensivists holding two conflicting defaults for the same patient. The MINT trial randomised 3,504 adults with acute myocardial infarction and anaemia to a restrictive threshold (transfuse below 7-8 g/dL) or a liberal one (below 10 g/dL), across 144 sites in six countries. This subanalysis asks whether the answer differs for the 1,679 patients (47.9%) who were in intensive care at randomisation.

It does not. The critically ill did far worse overall — 30-day death 11.9% against 6.5%, death or reinfarction 18.9% against 12.7% — but the effect of transfusion strategy tracked the same way in both groups. A restrictive strategy carried a non-significant 9% higher risk of death in the non-critically ill (risk ratio 1.09, 95% CI 0.77-1.54) and a non-significant 24% higher risk in the critically ill (1.24, 0.95-1.61), with interaction p values of 0.55 and 0.52.

So the overall MINT result stands for the ICU patient as much as for the ward patient, and the practical instruction is to stop treating the ICU as an exception. Where a patient has an acute myocardial infarction and anaemia, the threshold that applies is the one MINT supports — and the fact that they are ventilated or on vasopressors is not a reason to switch back to 7 g/dL.

• Apply the same transfusion threshold to an infarct patient in ICU as on the ward
• Remember the comparison is 7-8 g/dL against 10 g/dL, not transfusion against none
• Do not read the critically ill group's higher event rate as an effect of transfusion strategy
• Record haemoglobin trend and the threshold used, so the decision is reviewable
• In settings where blood supply is constrained, the small size of the difference matters to the decision

Why it matters

It removes a genuine conflict between two defaults that intensivists have been resolving by habit.

Don't overread it

This is a subgroup analysis and all its estimates are non-significant; it supports applying the main MINT result uniformly, not a claim that restrictive transfusion kills.

The statistics, in plain English

Every one of these risk ratios is non-significant, with intervals crossing 1.0 — so the honest reading is that this analysis cannot demonstrate a difference in either group, not that restrictive transfusion is proven harmful. The finding that matters is the interaction test: p values of 0.55 and 0.52 mean there is no evidence the two groups respond differently, which is what licenses applying the main trial result to ICU patients. Subgroup analyses are usually under-powered to detect interaction, so this is an absence of evidence, argued for consistency rather than for difference.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.