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, et al. (TRISS Trial Group)

Bottom Line

In patients with septic shock, a restrictive hemoglobin transfusion threshold of 7 g/dL did not result in a significant difference in 90-day mortality compared to a liberal threshold of 9 g/dL, while significantly reducing blood product utilization.

Key Findings

1. The primary outcome of 90-day mortality occurred in 216 of 502 patients (43.0%) in the restrictive group compared to 223 of 496 patients (45.0%) in the liberal group (relative risk 0.94; 95% confidence interval, 0.78-1.09; P=0.44).

2. Patients in the restrictive group received a median of 1 unit of red blood cells (interquartile range, 0-3), whereas those in the liberal group received a median of 4 units (interquartile range, 2-7) (P<0.001).

3. There were no significant differences between groups in secondary outcomes, including the occurrence of ischemic events (7.2% vs. 8.0%, P=0.64) or the requirement for life support at day 14.

4. Restrictive transfusion was shown to be safe, with no increase in adverse events or organ failure compared to the liberal strategy.

Study Design

Design

RCT

Assessor-blinded

Sample

998

Patients

Duration

90 days

Median

Setting

Multicenter, Scandinavia

Population Adult patients with septic shock and a hemoglobin concentration of 9 g/dL or less

Intervention Red blood cell transfusion at a hemoglobin threshold of 7 g/dL

Comparator Red blood cell transfusion at a hemoglobin threshold of 9 g/dL

Outcome Death by 90 days after randomization

Study Limitations

• The study excluded patients with acute coronary syndrome, which limits the applicability of the findings to septic shock patients with active cardiac ischemia.

• The trial was conducted exclusively in Scandinavian intensive care units, which may limit generalizability to healthcare systems with different transfusion practices or patient populations.

• The trial was not powered to detect small, yet potentially clinically important, differences in mortality, although it was adequately powered to test for the defined 9% absolute risk reduction.

Clinical Significance

The results of the TRISS trial support a restrictive transfusion strategy (hemoglobin trigger of 7 g/dL) as the standard of care for the majority of patients with septic shock, allowing for significant reductions in blood product usage without compromising patient safety or survival.

Historical Context

Prior to the TRISS trial, practice was heavily influenced by the early goal-directed therapy (EGDT) paradigm, which often encouraged higher hemoglobin thresholds (targeting >10 g/dL in early stages). The TRISS trial provided crucial evidence that challenged the need for liberal transfusion in sepsis, aligning with the earlier, seminal TRICC trial findings in the general ICU population.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

How does the relationship between hemoglobin concentration and systemic oxygen delivery ($DO_2$) explain why a lower transfusion threshold might be tolerated in patients with septic shock?

Key Response

Oxygen delivery ($DO_2$) is the product of cardiac output and arterial oxygen content ($CaO_2$). In septic shock, although $CaO_2$ decreases as hemoglobin drops, the body can often compensate by increasing cardiac output or increasing the oxygen extraction ratio ($OER$) at the tissue level. The TRISS trial demonstrates that for most patients, a hemoglobin of 7 g/dL provides sufficient $CaO_2$ to meet metabolic demands without the risks associated with blood transfusion, such as volume overload or transfusion-related lung injury.

Resident

A patient with septic shock also has a history of chronic stable angina. How should the TRISS trial results influence your transfusion trigger for this specific patient compared to a patient with no comorbidities?

Key Response

While the TRISS trial supported a restrictive threshold of 7 g/dL for the general septic shock population, it is important to note that patients with active myocardial ischemia were largely excluded or represent a subgroup where the benefit-risk ratio shifts. In patients with limited physiologic reserve or active coronary demand-supply mismatch, many clinicians and guidelines (including sub-analyses of restrictive vs. liberal trials) suggest a slightly higher threshold (e.g., 8 g/dL) to ensure adequate myocardial oxygenation, though the strict 9 g/dL threshold still showed no benefit in the overall TRISS cohort.

Fellow

The TRISS trial observed no difference in 90-day mortality between groups, but the liberal group received significantly more blood. Discuss the 'storage lesion' and immunomodulatory effects of RBC transfusion as potential reasons why increasing oxygen-carrying capacity did not improve outcomes in the liberal group.

Key Response

Stored red blood cells undergo biochemical and structural changes (the storage lesion), including reduced 2,3-DPG levels (shifting the oxyhemoglobin dissociation curve to the left) and decreased membrane deformability, which can impair microcirculatory flow—the very thing septic shock patients struggle with. Additionally, Transfusion-Related Immunomodulation (TRIM) can exacerbate the systemic inflammatory response in sepsis, potentially counteracting any benefit derived from a higher macrocirculatory hemoglobin level.

Attending

How did the TRISS trial fundamentally challenge the transfusion components of the original 2001 Early Goal-Directed Therapy (EGDT) protocol, and how should this affect our current bedside teaching?

Key Response

The original Rivers EGDT protocol recommended transfusing to a hematocrit of 30% (Hgb ~10 g/dL) if central venous oxygen saturation ($ScvO_2$) remained low. TRISS, following the TRICC trial's logic, proved that a trigger of 7 g/dL is safe even in the vasopressor-dependent phase of sepsis. This shifts our teaching from 'early aggressive' transfusion to 'physiologic tolerance,' emphasizing that $ScvO_2$ should be managed by optimizing flow and metabolic demand rather than simply increasing the RBC count.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

The TRISS trial used a 90-day mortality endpoint rather than a 28-day endpoint. What are the statistical and biological implications of this choice regarding the 'noise-to-signal' ratio in septic shock research?

Key Response

90-day mortality is increasingly preferred because it captures the 'long tail' of ICU mortality related to secondary infections and multi-organ failure. However, statistically, a longer follow-up introduces more confounding variables and non-attributable deaths (noise), which may dilute the observed treatment effect (signal) of an intervention delivered only during the acute phase. This necessitates a larger sample size to maintain power for detecting a significant difference in all-cause mortality.

Journal Editor

As a reviewer, what concerns would you raise regarding the 'open-label' design of the TRISS trial, and how might the lack of blinding influence secondary management that could affect the primary mortality endpoint?

Key Response

Blinding is impossible for transfusion triggers (clinicians see the lab results), which introduces 'co-intervention bias.' Clinicians aware that a patient is in the 'restrictive' group might be more cautious with fluid boluses or slower to wean vasopressors out of a perceived need to maintain perfusion at a lower Hgb. Conversely, they might be more aggressive with other interventions if they perceive the patient as 'under-treated.' While mortality is an objective endpoint, the pathway to it can be influenced by these non-standardized clinical decisions.

Guideline Committee

Based on the TRISS trial evidence, how should the Surviving Sepsis Campaign (SSC) grade the recommendation for a 7 g/dL threshold, and what specific exceptions must be codified in the guideline text?

Key Response

The TRISS trial provides high-quality evidence (Level A) to support a 'Strong Recommendation' for a restrictive transfusion strategy (threshold <7 g/dL) in adults with septic shock once hypoperfusion has resolved. However, the guidelines must explicitly state exceptions for patients with acute myocardial ischemia, severe hypoxemia, or acute hemorrhage, as these populations were either excluded from or underrepresented in the trial, and the physiological demands in these states differ significantly from distributive shock.

Clinical Landscape

Noteworthy Related Trials

1999

TRICC Trial

n = 838 · NEJM

Tested

Restrictive transfusion strategy (threshold 7.0 g/dL)

Population

Critically ill patients

Comparator

Liberal transfusion strategy (threshold 10.0 g/dL)

Endpoint

30-day all-cause mortality

Key result: A restrictive transfusion strategy was as effective as and possibly superior to a liberal transfusion strategy in critically ill patients, except perhaps for those with severe cardiac disease.

2011

FOCUS Trial

n = 2,016 · NEJM

Tested

Restrictive transfusion strategy (threshold 8.0 g/dL)

Population

Older patients with cardiovascular disease undergoing hip surgery

Comparator

Liberal transfusion strategy (threshold 10.0 g/dL)

Endpoint

60-day mortality or inability to walk 3 meters

Key result: A restrictive transfusion threshold did not increase the rate of death or inability to walk independently compared to a liberal threshold.

2015

TITRe2 Trial

n = 2,007 · NEJM

Tested

Restrictive transfusion strategy (threshold 7.0-8.0 g/dL)

Population

Patients undergoing cardiac surgery

Comparator

Liberal transfusion strategy (threshold 9.0 g/dL)

Endpoint

Infection or ischemic events at 90 days

Key result: A restrictive transfusion strategy was not shown to be superior to a liberal strategy in reducing rates of infection or ischemic events.

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