A restrictive RBC transfusion strategy is preferred for most critically ill adults (including GI bleeding and cardiac surgery) because it reduces transfusions without increasing mortality or complications; avoid a restrictive strategy in acute coronary syndrome (ACS).

Background

About one-quarter of ICU patients receive red blood cell (RBC) transfusions, most commonly for anemia. Transfusions carry risks (eg, transfusion-related acute lung injury, circulatory overload, infections) and substantial costs. Given new trials and practice variability, the American College of Chest Physicians (CHEST) panel synthesized contemporary evidence to guide transfusion thresholds in critically ill adults overall and in key subgroups.

Patients

• Critically ill adults in ICU settings (excluding active, unstable hemorrhage; neurologic injury; and trauma).
• Subgroups: acute gastrointestinal bleeding (GIB), acute coronary syndrome (ACS), cardiac surgery, isolated troponin elevation (without clinical ischemia), and septic shock with end-organ hypoperfusion.
• Evidence base: 23 studies (22 randomized controlled trials and 1 cohort); overall critically ill population plus subgroup trials.

Intervention

Restrictive RBC transfusion strategy (typically transfuse at hemoglobin 7–8 g/dL; transfuse one unit at a time and recheck hemoglobin).

Control

Permissive (liberal) RBC transfusion strategy (typically transfuse at hemoglobin 8.5–10 g/dL), or adding permissive transfusion thresholds to usual care in septic shock.

Outcome

• Mortality (ICU, 30-day, 6-week, and 1-year).
• Adverse events: transfusion reactions; infections; cardiac, renal, pulmonary, thromboembolic complications; rebleeding (GIB); need for surgery or revascularization (as applicable).
• Resource use: number of units transfused; ICU and hospital length of stay (LOS); costs.

Study Design

CHEST clinical practice guideline using systematic review and meta-analysis; GRADE framework for certainty and recommendation strength; modified Delphi for consensus.

Level of Evidence

• Overall critically ill and GIB: moderate certainty; strong recommendations for restrictive strategy.
• Cardiac surgery: moderate certainty; conditional for restrictive strategy.
• ACS: low certainty; conditional against restrictive strategy.
• Isolated troponin elevation: very low certainty; conditional for restrictive strategy.
• Septic shock: low certainty; conditional against adding permissive thresholds to usual care.

Follow up period

• Short-term: ICU and 30-day mortality widely reported; 6-week mortality in upper GIB.
• Long-term: 1-year mortality reported in ACS and septic shock cohorts.

Results

Overall critically ill adults

Primary outcomes

• Mortality: No difference between restrictive and permissive strategies for ICU mortality (RR 1.00; 95% CI 0.80–1.25), 30-day mortality (RR 0.99; 95% CI 0.87–1.13), or 1-year mortality (RR 0.99; 95% CI 0.87–1.13).

Secondary outcomes

• Adverse events (overall): Lower with restrictive strategy (RR 0.45; 95% CI 0.22–0.94); absolute risk reduction 8 per 1,000 → NNT ≈ 125.
• Infections, organ-specific complications, thromboembolism, LOS: No clinically meaningful differences. ICU LOS difference (~+0.12 days) not clinically significant.
• Resource use: Restrictive strategy reduced the number of RBC units transfused by ~50%.

Acute gastrointestinal bleeding

Primary outcomes

• 30-day (6-week) mortality: Restrictive superior (RR 0.68; 95% CI 0.48–0.97); absolute risk reduction 27 per 1,000 → NNT ≈ 37.

Secondary outcomes

• Acute transfusion reactions: Reduced (RR 0.35; 95% CI 0.20–0.61); absolute reduction 37 per 1,000 → NNT ≈ 27.
• Serious adverse transfusion effects: Reduced (RR 0.73; 95% CI 0.58–0.91); absolute reduction 54 per 1,000 → NNT ≈ 19.
• Rebleeding-related surgery, infections, organ-specific complications, hospital LOS: No significant differences.

Acute coronary syndrome

Primary outcomes

• 30-day mortality: Trend favoring permissive strategy; pooled RR 1.13 (95% CI 0.67–1.91) for restrictive versus permissive (no significant difference). Large MINT trial showed higher cardiac death with restrictive threshold.

Secondary outcomes

• MI or death, recurrent MI: Point estimates favor permissive strategy.
• Infections, thromboembolism, organ-specific complications, LOS: No clear differences.

Implication: Avoid a restrictive threshold in ACS; consider higher transfusion thresholds (often 9–10 g/dL) individualized to symptoms/physiology.

Cardiac surgery (perioperative)

Primary outcomes

• 30-day mortality: No difference (RR 1.12; 95% CI 0.95–1.32) between restrictive and permissive strategies.

Secondary outcomes

• ICU/hospital LOS, infections, cardiac/renal/pulmonary complications, thromboembolism: No important differences.
• Resource use: Fewer patients transfused and fewer units with restrictive thresholds.

Isolated troponin elevation (without clinical ischemia)

• No direct trial data; panel suggests a restrictive approach with individualized decisions considering comorbidity, hemodynamics, and indicators of perfusion/oxygen delivery.

Septic shock with end-organ hypoperfusion

Primary outcomes

• 30-day mortality and ICU mortality: No differences between adding permissive thresholds and restrictive thresholds (eg, RR 0.93; 95% CI 0.72–1.21 for 30-day mortality).

Secondary outcomes

• Transfusion reactions, renal replacement therapy, cardiac and pulmonary complications, LOS: No significant differences.
• Resource use: Restrictive thresholds reduce RBC utilization and cost.

Limitations

• Certainty ranged from very low to moderate; some outcomes imprecise and underpowered (eg, transfusion reactions).
• Heterogeneity in thresholds (restrictive 7–8 g/dL; permissive 8.5–10 g/dL), populations, and settings.
• Limited data in ACS (uncertain optimal threshold between 9 and 10 g/dL), septic shock subpopulations, and no direct data for isolated troponin elevation without ischemia.
• Excluded protocolized early goal-directed resuscitation trials; did not address RBC storage age, processing, or donor factors.
• No recommendations for patients with active uncontrolled hemorrhage, neurologic injury, or trauma (studies ongoing).

Funding

No external funding reported.

Citation

Coz Yataco AO, Soghier I, Hébert PC, Belley-Cote E, Disselkamp M, et al. Red Blood Cell Transfusion in Critically Ill Adults: An American College of Chest Physicians Clinical Practice Guideline. CHEST. 2025;167(2):477-489. doi:10.1016/j.chest.2024.09.016