Skip to main content

Foundational Trials

TRICC

Hébert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340(6):409‑417

Image: Shutterstock / Chan2545

Paper

Sections

Publication  |  Context & Rationale  |  Design & Methods  |  Key Results  |  Internal Validity  |  External Validity  |  Strengths & Limitations  |  Interpretation & Why It Matters  |  Controversies & Subsequent Evidence  |  Summary  |  Further Reading  |  Overall Takeaway

Publication

  • Title

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

  • Acronym
    TRICC
  • Year

    1999

  • Citation

    Hébert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340(6):409‑417

Context & Rationale

  • Background
    • By the mid‑1990s most ICU protocols reflexively transfused when hemoglobin fell below 10 g/dL (“10/30 rule”), despite observational data suggesting transfusion‑related immunosuppression, lung injury, infection, and limited survival benefit.
    • No large RCT had tested whether lower thresholds were safe.
  • Why This Matters
    • Blood is costly and finite; avoiding unnecessary transfusions reduces exposure to infectious, immunologic, and circulatory complications.
    • The trial therefore tested a practice‑changing question with major resource and safety implications.
  • Question

    Does a restrictive strategy (transfuse at Hb < 7 g/dL, maintain 7–9 g/dL) achieve equivalent or better outcomes than a liberal strategy (transfuse at Hb < 10 g/dL, maintain 10–12 g/dL) in stable, critically ill adults?

Design & Methodology

Trial Design

  • Design

    Prospective, randomized, concealed‑allocation, unblinded, multicenter equivalence trial

  • Setting

    22 tertiary and 3 community Canadian ICUs

  • Period

    November 1994 to November 1997

Population

  • Inclusion Criteria
    • Expected ICU stay > 24 h
    • Hb ≤ 9 g/dL within 72 h of admission
    • judged euvolemic
  • Exclusion Criteria
    • Age < 16 y
    • active bleeding (≥ 3 U/12 h or Hb drop ≥ 3 g/dL)
    • chronic anemia
    • pregnancy
    • brain death/imminent death
    • DNR/ treatment‑withdrawal discussions
    • routine postoperative cardiac surgery
    • inability to receive blood products

Intervention

  • Trigger

    Transfuse one unit of packed red blood cells (PRBC) unit if Hb < 7 g/dL

  • Target

    Target 7–9 g/dL

Control

  • Trigger

    Transfuse if Hb < 10 g/dL

  • Target

    Target 10–12 g/dL

Statistical Plan

  • Power Calculation

    Original equivalence design required 2,300 patients to rule out a 4 % absolute difference (α = 0.05, β = 0.20)

  • Interim Analysis

    After an interim analysis showed higher mortality than predicted, the sample size was reduced to 1,620

  • Analysis

    Intention-to-treat

Other

  • Blinding
    • Open‑label (hemoglobin values drive transfusion decisions).
    • Primary outcomes (mortality) - objective
    • Organ‑dysfunction scoring unblinded but protocolised
  • Follow Up
    • ICU stay (protocol exposure)
    • Hospital discharge
    • 30‑day (primary) 
    • 60‑day mortality

Key Results

Restrictive versus Liberal Groups

  • Early Stopping
    • The trial was stopped early for poor enrolment, which had slowed to <20% predicted
    • The trial was not stopped for efficacy.

Primary Outcome

  • 30‑day mortality
    • 78 / 418 (18.7 %) vs 98 / 420 (23.3 %)
    • RR 0.80
    • P=0.11

Secondary Outcomes

  • In‑hospital mortality
    • 93 / 418 (22.2 %) vs 118 / 420 (28.1 %)
    • RR 0.79
    • P = 0.05
  • Mean units transfused
    • 2.6 ± 4.1 vs 5.6 ± 5.3
    • Δ –53 %
    • P<0.01
  • Any ICU complication
    • 205 (49 %) vs 228 (54 %)
    • Δ −5 %
    • 0.12

  • Myocardial infarction
    • 3 (0.7 %) vs 12 (2.9 %)
    • RR 0.24
    • P=0.02

Notes

  • Mortality was numerically lower with the restrictive strategy and significantly lower in prespecified subgroups:
    • age < 55 y (5.7 % vs 13.0 %)
    • APACHE II ≤ 20 (8.7 % vs 16.1 %)
  • Cardiac complications (MI, pulmonary edema) were more common with liberal transfusion

Internal Validity

  • Randomisation & Allocation
    • Central, block‑randomization
    • Sealed opaque envelopes
    • Stratified by:
      • Centre
      • APACHE II
  • Performance/Detection Bias
    • Open‑label design unavoidable
    • Primary outcome objective
  • Protocol Adherence
    • Non‑compliance < 5 %
    • Cross‑over 1.8 %
  • Outcome Assessment
    • Mortality hard end‑point
    • Organ‑dysfunction scores standardised
  • Statistical Rigor
    • Equivalence framework appropriate
    • Early stoppage reduced power but the point estimate favoured the restrictive arm
  • Separation of the Variable of Interest
    • Hemoglobin separation (8.5 vs 10.7 g/dL)
    • 54 % fewer units demonstrate strong protocol adherence
  • Key Delivery Aspects
    • Transfusion decisions triggered promptly by measured Hb
    • Clear separation achieved
  • Attrition
    • Minimal risk of attrition bias
      • 1 % withdrew
      • <0.5 % lost to follow‑up

  • Baseline Characteristics

    Groups well matched by:

    • Age (≈ 58 y)
    • APACHE II ≈ 21
    • Organ‑failure scores
    • Diagnosis mix
  • Heterogeneity
    • Broad case‑mix improves generality
    • Subgroup analyses pre‑specified
  • Conclusion

    Moderately strong—methodology robust, but open‑label design and premature closure introduce some uncertainty

External Validity

  • Population Representativeness
    • Adults in mixed Canadian ICUs with stable hemodynamics;
    • Important exclusions included
      • Active bleeding
      • Chronic anemia
      • Immediate post‑cardiac‑surgery patients
  • Applicability
    • Findings generalisable to most haemodynamically stable ICU patients in high‑income systems
    • Caution in acute coronary syndromes (signal of harm not excluded)
  • Conclusion
    • Good for general ICU practice
    • Limited for
      • Actively bleeding
      • Ischemic heart disease populations

Strengths & Limitations

Strengths

  • Design

    Multicenter pragmatic design with concealed allocation

  • Group separation

    Clear separation of hemoglobin and transfusion exposure

  • Subgroups

    Pre‑specified clinically relevant subgroups

Limitations

  • Open label

    Potential co‑intervention bias

  • Underpowered

    Underpowered due to early termination at 52 % of revised target

  • Ischaemia

    Excluded unstable cardiac ischemia, leaving an unanswered risk in that patient group

Interpretation / Why This Matters

  • Landmark trial

    TRICC showed that a hemoglobin threshold of 7 g/dL is at least as safe as 10 g/dL and often advantageous, cutting transfusion exposure by half without increasing—and possibly lowering—mortality and cardiac events.

  • Paradigm Changing

    It overturned the traditional “10/30” dogma and paved the way for restrictive‑threshold recommendations worldwide

Controversies & Subsequent Evidence

  • Statistical significance

    Critics argued the non‑significant primary end‑point reflected low power; however, point estimates and subgroup benefits persisted in later meta‑analyses

  • Stable coronary disease

    A post‑hoc analysis (Crit Care Med 2001) showed neutral outcomes in patients with stable cardiac disease, fueling debate about thresholds in ischemia

  • Subsequent support

    TRIPICU (pediatrics), FOCUS (hip fracture), TRISS (septic shock) and most modern ICU/operative trials corroborate safety of 7–8 g/dL thresholds

  • Guidelines

    The 2012 and 2016 AABB guidelines endorse restrictive strategies (7–8 g/dL) for stable hospitalised patients and many cardiovascular patients, citing TRICC as foundational.

Summary

  • Reduced transfusion

    Restrictive (7 g/dL) transfusion reduced PRBC use by 54 % yet yielded similar or lower mortality

  • Benefits without obvious harm

    Benefits most evident in younger, less‑severely ill subgroups; no harm detected overall

  • Possible increase in cardiac complications

    Liberal transfusion increased cardiac complications

  • Early stopping

    Early termination limits certainty but subsequent trials confirm findings

  • Basis for current practice

    Practice‑changing evidence underpinning modern restrictive transfusion guidelines

Conclusion

  • Landmark Trial
    • TRICC changed the practice of anaemia management worldwide, a practice that has lasted 25 years.
    • However, the recent triumpherate of neuro transfusion trials (Hemotion, TRAIN & SAHaRA) now somewhat challenge this restrictive practice in the neuro ICU, as have MINT and REALITY in acute myocardial infarction.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Overall Takeaway

TRICC decisively challenged the long‑standing liberal transfusion paradigm, demonstrating that “less is more” for most critically ill adults. Its findings reshaped transfusion practice, informed international guidelines, and continue to anchor modern restrictive thresholds in ICU and peri‑operative care.

Last updated August 19th, 2025