The New England Journal of Medicine FEBRUARY 11, 1999

A Multicenter, Randomized, Controlled Clinical Trial of Transfusion Requirements in Critical Care (TRICC)

Hébert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G, Tweeddale M, Schweitzer I, Yetisir E

Bottom Line

The TRICC trial established that a restrictive red-cell transfusion strategy, targeting a hemoglobin concentration of 7.0–9.0 g/dL, is at least as effective as, and potentially superior to, a liberal strategy targeting 10.0–12.0 g/dL in critically ill patients, with the exception of those with active cardiac ischemia.

Key Findings

1. Overall 30-day mortality was not statistically significantly different between the restrictive (18.7%) and liberal (23.3%) groups (P=0.11).

2. Hospital mortality was significantly lower in the restrictive-strategy group compared to the liberal-strategy group (22.2% vs. 28.1%, P=0.05).

3. In subgroup analyses, restrictive transfusion was associated with significantly lower 30-day mortality in patients with an APACHE II score ≤20 (8.7% vs. 16.1%, P=0.03) and patients <55 years of age (5.7% vs. 13.0%, P=0.02).

4. Cardiac complications (e.g., pulmonary edema and myocardial infarction) were more common in the liberal transfusion group (21.0%) than in the restrictive group (13.2%, P<0.01).

5. No significant difference in mortality was observed in patients with clinically significant cardiac disease (20.5% vs. 22.9%, P=0.69).

Study Design

Design

RCT

Open-Label

Sample

838

Patients

Duration

30 days

Median

Setting

Multicenter, Canada

Population Critically ill adults expected to stay in the ICU >24 hours with euvolemia after initial treatment and hemoglobin ≤9.0 g/dL within 72 hours of admission.

Intervention Restrictive transfusion strategy: red cells transfused if hemoglobin dropped below 7.0 g/dL, with a target of 7.0–9.0 g/dL.

Comparator Liberal transfusion strategy: red cells transfused if hemoglobin dropped below 10.0 g/dL, with a target of 10.0–12.0 g/dL.

Outcome All-cause mortality at 30 days.

Study Limitations

• The trial was terminated prematurely for slow enrollment, which reduced the statistical power to demonstrate equivalence.

• The study was unblinded (open-label), which may introduce bias despite the objective nature of the primary mortality endpoint.

• Exclusion criteria were stringent, potentially limiting generalizability to patients with chronic anemia, acute active hemorrhage, or those requiring cardiac surgery.

• The trial did not specifically power for or include enough patients with active myocardial ischemia to draw definitive conclusions for that high-risk subgroup.

Clinical Significance

The TRICC trial provided the critical evidentiary basis to challenge the long-standing '10/30' transfusion rule (maintaining hemoglobin at 10 g/dL). It demonstrated that lower thresholds are safe and possibly beneficial for most critically ill adults, encouraging more judicious blood product utilization and reducing the morbidity associated with unnecessary transfusions.

Historical Context

Prior to 1999, the standard of care for critically ill patients was to maintain hemoglobin levels at approximately 10 g/dL (the '10/30' rule), primarily based on physiological assumptions about oxygen delivery. The TRICC trial was a pioneering randomized study that shifted the paradigm toward restrictive transfusion thresholds in the intensive care setting.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

How does the oxygen delivery (DO2) equation explain why a lower hemoglobin concentration might be tolerated in a critically ill patient without causing immediate tissue hypoxia?

Key Response

DO2 is the product of Cardiac Output (CO) and Arterial Oxygen Content (CaO2), where CaO2 = (Hb x 1.34 x SaO2) + (PaO2 x 0.003). In the TRICC trial, patients with a lower hemoglobin (restrictive group) maintained tissue perfusion because the body compensates for reduced oxygen-carrying capacity (Hb) by increasing cardiac output and increasing the oxygen extraction ratio (O2ER). Understanding this basic physiology is essential for clinical reasoning regarding when a transfusion is truly 'necessary' versus 'elective'.

Resident

The TRICC trial suggested improved outcomes with a restrictive strategy in certain subgroups; which specific patient populations showed a trend toward higher mortality when a liberal transfusion strategy was employed?

Key Response

Subgroup analysis in the TRICC trial revealed that younger patients (aged <55 years) and those who were less acutely ill (APACHE II score <= 20) had significantly lower mortality rates with the restrictive strategy. This implies that for patients with more physiological reserve, the risks of transfusion (such as TRALI, TACO, and immunomodulation) may outweigh the benefits of an increased hemoglobin level, making the restrictive threshold (7.0 g/dL) the standard of care for these groups.

Fellow

In light of the TRICC trial's results, how should the presence of active myocardial ischemia or unstable angina alter your application of transfusion thresholds in the ICU setting compared to a patient with sepsis?

Key Response

The TRICC trial excluded patients with active myocardial ischemia and noted that mortality was actually lower (though not significantly) in the liberal group for those with known ischemic heart disease. Subsequent trials like REALITY and MINT have further explored this, suggesting that while 7 g/dL is safe for sepsis (as confirmed by the TRISS trial), a higher threshold of 8-9 g/dL is often more appropriate for those with impaired coronary oxygen delivery, requiring fellows to integrate trial-specific exclusion criteria into broad clinical protocols.

Attending

Reflecting on the 'storage lesion' of red blood cells, how does the age of transfused blood potentially provide a mechanistic explanation for the TRICC trial's findings that a liberal strategy was not superior?

Key Response

Stored RBCs undergo biochemical and structural changes, including depletion of 2,3-DPG (shifting the oxygen-dissociation curve left) and decreased deformability. In a liberal strategy, patients receive more units of potentially 'older' blood, which can cause microvascular obstruction and release pro-inflammatory cytokines. Teaching this highlights that 'more blood' does not always equate to 'more oxygen delivery at the tissue level,' justifying the restrictive approach as a means of 'doing no harm'.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

The TRICC trial utilized a 2.0 g/dL 'buffer zone' between the restrictive (7-9 g/dL) and liberal (10-12 g/dL) arms; how does this gap affect the statistical power of the study and the ability to define an 'optimal' physiological hemoglobin set-point?

Key Response

By creating a distinct separation between arms, the researchers maximized the likelihood of observing a treatment effect (discriminatory power). However, this 'gap' design makes it difficult to determine the safety of intermediate values (e.g., 9.5 g/dL). A researcher would critique this as a trade-off between internal validity (ensuring the groups are truly different) and the ability to model the dose-response relationship between hemoglobin and mortality.

Journal Editor

As an editor, how would you evaluate the threat to external validity posed by the fact that only 838 patients were randomized out of over 8,000 screened in the TRICC trial?

Key Response

A screening-to-enrollment ratio of approximately 10% is a significant red flag for selection bias. It suggests that clinicians were only comfortable randomizing 'stable' patients and excluded those they felt 'needed' blood (physician preference). This limits the generalizability of the results to the most critically unstable ICU patients. An editor would require a transparent 'CONSORT' flow diagram and a robust discussion of how the 'refusal to randomize' might have skewed the study population toward a lower-risk cohort.

Guideline Committee

How do the TRICC results compare with current AABB and SCCM guidelines, and what specific grade of recommendation is now assigned to the restrictive transfusion threshold for hemodynamically stable ICU patients?

Key Response

Current guidelines (e.g., AABB 2023) provide a 'Strong Recommendation' with 'High-Certainty Evidence' (Grade 1A) for a restrictive threshold of 7 g/dL for most hospitalized, hemodynamically stable adult patients, including those in the ICU. This is a direct result of TRICC's foundational work, which has been validated by subsequent meta-analyses. The committee must reconcile this with the weaker 'conditional' recommendations for patients with acute myocardial infarction (threshold usually 8 g/dL), reflecting the ongoing nuances in the evidence base.

Clinical Landscape

Noteworthy Related Trials

2011

FOCUS Trial

n = 2,016 · NEJM

Tested

Liberal transfusion strategy (Hb >10 g/dL)

Population

Older patients with cardiovascular disease undergoing hip surgery

Comparator

Restrictive transfusion strategy (Hb >8 g/dL)

Endpoint

Death or inability to walk independently at 60 days

Key result: The restrictive strategy was not inferior to the liberal strategy regarding the primary outcome.

2014

TRISS Trial

n = 1,005 · NEJM

Tested

Restrictive transfusion strategy (Hb <7 g/dL)

Population

Patients with septic shock in the ICU

Comparator

Liberal transfusion strategy (Hb <9 g/dL)

Endpoint

Death at 90 days

Key result: The restrictive transfusion threshold resulted in fewer transfusions but did not affect mortality compared to the liberal strategy.

2015

TITRE2 Trial

n = 2,007 · NEJM

Tested

Restrictive transfusion strategy (Hb <7.5 g/dL)

Population

Patients undergoing cardiac surgery

Comparator

Liberal transfusion strategy (Hb <9 g/dL)

Endpoint

Infection or ischemic events at 90 days

Key result: Restrictive transfusion threshold did not result in higher rates of infection or ischemic events compared to liberal transfusion.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis