New England Journal of Medicine October 09, 2014

Lower versus Higher Hemoglobin Threshold for Transfusion in Septic Shock

Lars B. Holst, Nicolai Haase, Jørn Wetterslev, Jan Wernerman, Anne B. Guttormsen, Sari Karlsson, Pär I. Johansson, Anders Åneman, et al.

Source: View publication →

Bottom Line

In patients with septic shock, a restrictive hemoglobin threshold for red blood cell transfusion (7 g/dL) resulted in similar 90-day mortality and ischemic event rates compared to a liberal threshold (9 g/dL), while significantly reducing the number of transfusions.

Key Findings

1. At 90 days, mortality was similar between the restrictive-threshold group and the liberal-threshold group (43.0% vs. 45.0%; Relative Risk 0.94, 95% CI 0.78 to 1.09, P=0.44).

2. Patients in the restrictive group received significantly fewer red blood cell transfusions in the ICU, with a median of 1 unit (interquartile range, 0 to 3) compared to 4 units (interquartile range, 2 to 7) in the liberal group.

3. The rates of ischemic events, severe adverse reactions, and the need for life support (vasopressors, mechanical ventilation, and renal replacement therapy) did not differ significantly between the two intervention groups.

Study Design

Design

Randomized Controlled Trial

Open-Label

Sample

998

Patients

Duration

90 days

Median

Setting

Multicenter, Scandinavia

Population Adult ICU patients with septic shock and a baseline hemoglobin concentration of 9.0 g/dL or less.

Intervention Restrictive transfusion strategy: 1 unit of leukoreduced red blood cells transfused only when the hemoglobin level was 7.0 g/dL or less.

Comparator Liberal transfusion strategy: 1 unit of leukoreduced red blood cells transfused when the hemoglobin level was 9.0 g/dL or less.

Outcome Death from any cause by 90 days after randomization.

Study Limitations

• The open-label design for clinical staff could have introduced management bias, although the objective primary outcome of 90-day mortality mitigates this risk.

• Patients with acute coronary syndrome or active bleeding were excluded, limiting the applicability of a strict 7 g/dL threshold in these high-risk subgroups.

• The baseline overall mortality in the trial was high (~44%), representing a very sick patient population; findings may be extrapolated differently in lower-acuity settings.

Clinical Significance

The TRISS trial established that a restrictive transfusion strategy (hemoglobin < 7 g/dL) is safe and effective for patients with septic shock. This challenged the prior emphasis on higher hemoglobin targets (e.g., maintaining hematocrit > 30%) derived from Early Goal-Directed Therapy (EGDT) protocols, ultimately sparing patients from unnecessary blood product exposure, mitigating transfusion-related adverse events, and conserving healthcare resources.

Historical Context

The 1999 TRICC trial demonstrated the safety of a restrictive transfusion threshold in general critical care. However, the 2001 Rivers Early Goal-Directed Therapy (EGDT) trial recommended a liberal threshold (hematocrit > 30%) specifically for early sepsis to optimize central venous oxygen saturation (ScvO2). This led to conflicting guidelines in critical care medicine until TRISS definitively showed in 2014 that the restrictive threshold was equally safe in the septic shock population.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

Based on the oxygen delivery equation, why might one hypothesize that a higher hemoglobin threshold would benefit patients in septic shock, and physiologically, why did the TRISS trial show no difference in mortality?

Key Response

Tests knowledge of DO2 = CO x (1.34 x Hb x SaO2 + PaO2 x 0.003). It highlights that while higher Hb theoretically increases DO2, septic shock involves microcirculatory shunting and mitochondrial dysfunction. Thus, extra RBCs do not necessarily improve cellular dysoxia, making a restrictive strategy safer and avoiding transfusion-related harms.

Resident

A patient with septic shock and a history of stable coronary artery disease has a hemoglobin of 7.5 g/dL. According to the TRISS trial, should this patient be transfused, and how does this approach differ from the older early goal-directed therapy (EGDT) protocols?

Key Response

Explores the shift from EGDT (which targeted Hct >30% if ScvO2 <70%) to modern restrictive strategies (Hb < 7). TRISS showed safety in restrictive transfusion even in stable CAD, teaching residents to reserve transfusions for Hb < 7 unless active acute coronary syndrome or severe hypoxemia is present.

Fellow

The TRISS trial found no difference in ischemic events between restrictive and liberal groups. How do the physiological effects of banked blood potentially offset the theoretical benefits of a higher hemoglobin target in the microcirculation of a profoundly vasoplegic patient?

Key Response

Dives deep into the storage lesion of transfused RBCs, such as the loss of 2,3-DPG (shifting the oxyhemoglobin dissociation curve left) and reduced RBC deformability (impairing microvascular transit), which explains why empirical liberal transfusion fails to improve tissue oxygenation in sepsis.

Attending

Given the robust findings of TRISS alongside TRICC, what are the most common cognitive biases that persist among clinicians when a septic patient with an Hb of 7.1 g/dL appears poorly perfused, and how can we mitigate this transfusion trigger reflex in clinical practice?

Key Response

Addresses the gap between evidence and practice. Attendings must teach that tachycardia and lactatemia in sepsis are often due to distributive physiology and mitochondrial dysfunction, not absolute oxygen-carrying capacity, and that unnecessary transfusions carry risks like TACO, TRALI, and immunomodulation without outcome benefit.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

The TRISS trial used a pragmatic, unblinded design for the intervention but blinded the outcome assessors. What are the specific methodological risks of performance bias in unblinded ICU transfusion trials, and how did investigators handle the protocol violations statistically?

Key Response

Critiques the unblinded nature of the intervention where clinicians knowing the group might alter co-interventions like fluids or pressors. Discusses per-protocol versus intention-to-treat analyses and the handling of crossover or non-adherence in critical care trials.

Journal Editor

How does the exclusion of patients with acute myocardial ischemia threaten the external validity of the safe-to-restrict conclusion, and was the mean 1.5 g/dL difference in hemoglobin achieved between arms sufficient to definitively reject a benefit of liberal transfusion?

Key Response

Evaluates the generalizability boundary, as active ischemia was excluded, making the 7 g/dL threshold questionable for that specific subset. It also challenges whether the physiological separation was wide enough to test the hypothesis robustly without risking a Type II error in specific subgroups.

Guideline Committee

How does the evidence from the TRISS trial directly inform the Surviving Sepsis Campaign (SSC) guidelines regarding red blood cell transfusion, and what specific strength of recommendation and quality of evidence is assigned to the 7.0 g/dL threshold?

Key Response

Links TRISS directly to SSC guidelines, which strongly recommend a restrictive strategy (Hb < 7 g/dL) in adults with septic shock (Strong recommendation, high quality of evidence), explicitly moving away from EGDT targets unless specific exceptions like active myocardial ischemia or severe hemorrhage exist.

Clinical Landscape

Noteworthy Related Trials

1999

TRICC Trial

n = 838 · NEJM

Tested

Restrictive transfusion strategy (Hb threshold <7.0 g/dL)

Population

Critically ill ICU patients

Comparator

Liberal transfusion strategy (Hb threshold <10.0 g/dL)

Endpoint

30-day all-cause mortality

Key result: A restrictive strategy of red-cell transfusion is at least as effective as and possibly superior to a liberal transfusion strategy in critically ill patients.

2001

Rivers EGDT Trial

n = 263 · NEJM

Tested

Early goal-directed therapy including transfusion to hematocrit >30% for low ScvO2

Population

Patients with severe sepsis or septic shock

Comparator

Standard therapy

Endpoint

In-hospital mortality

Key result: EGDT significantly reduced in-hospital mortality compared to standard therapy.

2014

ProCESS Trial

n = 1341 · NEJM

Tested

Protocol-based EGDT (requiring central venous monitoring and transfusion targets)

Population

Emergency department patients with early septic shock

Comparator

Protocol-based standard care or usual care

Endpoint

60-day in-hospital mortality

Key result: Protocol-based resuscitation with strict targets like hematocrit >30% did not improve outcomes over standard care.

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