← Back

SAHARA

Liberal or Restrictive Transfusion Strategy in Aneurysmal Subarachnoid Hemorrhage

Year of Publication: 2025

Authors: S.W. English, A. Delaney, D.A. Fergusson, ..., for the SAHARA Trial Investigators on behalf of the Canadian Critical Care Trials Group*

Journal: N Engl J Med

Citation: N Engl J Med 2025;392:1079-88. DOI: 10.1056/NEJMoa2410962

Link: https://www.nejm.org/doi/full/10.1056/NEJMoa2410962

PDF: https://www.researchgate.net/publication...mxpY2F0aW9uIn19

Clinical Question

Does a liberal red-cell transfusion strategy, as compared with a restrictive strategy, result in a lower risk of an unfavorable neurologic outcome at 12 months in critically ill adults with acute aneurysmal subarachnoid hemorrhage and anemia?

Bottom Line

In patients with aneurysmal subarachnoid hemorrhage and anemia, a liberal transfusion strategy did not result in a lower risk of an unfavorable neurologic outcome at 12 months than a restrictive strategy. The incidence of adverse events was similar between the two groups.

        Major Points
        The SAHARA trial was an investigator-initiated, pragmatic, open-label, randomized trial with a blinded outcome assessment.
A total of 742 patients were randomized at 23 centers across Canada, Australia, and the United States.
The primary outcome was an unfavorable neurologic outcome (modified Rankin scale score ≥4) at 12 months.
An unfavorable neurologic outcome occurred in 33.5% of the liberal-strategy group and 37.7% of the restrictive-strategy group (risk ratio, 0.88; 95% CI, 0.72 to 1.09; P=0.22).
Secondary outcomes, including functional independence (FIM score), quality of life (EQ-5D-5L utility index, VAS score), and mortality, did not show significant differences between groups.
Radiographic vasospasm was observed in 31.5% of the liberal-strategy group and 40.7% of the restrictive-strategy group (risk ratio, 0.77; 95% CI, 0.63 to 0.95).
The incidence of delayed cerebral ischemia and new cerebral infarction was similar in both groups.
Red-cell transfusions were received by 99.7% of the liberal-strategy group and 34.9% of the restrictive-strategy group.

      

Design

Study Type: Randomized Controlled Trial

Randomization: 1

Blinding: Blinded outcome assessment by independent assessors who were unaware of trial-group assignments.

Enrollment Period: Pilot trial: October 17, 2015, to November 21, 2016; Full trial: March 11, 2018, to July 10, 2023.

Follow-up Duration: 12 months

Centers: 23

Countries: Canada, Australia, United States

Sample Size: 742

Analysis: Intention-to-treat. Primary outcome analyzed using a binomial generalized linear mixed model with fixed effects for treatment and sex and a random effect for center (log-link function). Proportional-odds analysis also conducted. Secondary outcomes analyzed using linear generalized linear mixed model with an identity-link function. SAS software, version 9.4; R software, version 4.3.1.

Inclusion Criteria

Patients 18 years of age or older.
First-ever aneurysmal subarachnoid hemorrhage.
Hemoglobin level of ≤10 g per deciliter.
Within the first 10 days after admission (pilot trial: within first 14 days).

Exclusion Criteria

Nonaneurysmal subarachnoid hemorrhage.
Active bleeding leading to hemodynamic instability.
Contraindications to blood transfusion.
Known objections to blood transfusion.

Baseline Characteristics

Characteristic	Control	Active
Age - yr Mean	59.3±12.5	59.5±12.2
Age - yr Median (IQR)	60 (51-68)	60 (52-68)
Sex - Female no. (%)	299 (81.7)	299 (81.7)
Sex - Male no. (%)	67 (18.3)	67 (18.3)
Hypertension no. (%)	161 (44.0)	177 (48.4)
Current or previous smoker no. (%)	123 (33.6)	149 (40.7)
Ischemic heart disease no. (%)	3 (0.8)	3 (0.8)
Congestive heart failure no. (%)	15 (4.1)	28 (7.7)
History of intracranial hemorrhage no. (%)	1 (0.3)	5 (1.4)
Ischemic stroke no. (%)	7 (1.9)	14 (3.8)
End-stage renal disease no. (%)	1 (0.3)	0
Chronic anemia no. (%)	20 (5.5)	23 (6.3)
No. of days from first hospital presentation to randomization Mean	3.6±2.2	3.8±2.5
No. of days from first hospital presentation to randomization Median (IQR)	3 (2-5)	3 (2-5)
Median score on the World Federation Neurological Surgeons scale at admission (IQR)	3 (1-4)	3 (1-4)
Median score on the modified Fisher scale (IQR)	4 (3-4)	4 (3-4)
Externalized ventricular drain no. (%)	257 (70.2)	255 (69.7)
Location of aneurysm - Anterior circulation no. (%)	328 (89.6)	310 (84.7)
Location of aneurysm - Posterior circulation no. (%)	38 (10.4)	56 (15.3)
Aneurysm treatment - Endovascular no. (%)	218 (59.6)	225 (61.5)
Aneurysm treatment - Neurosurgical no. (%)	151 (41.3)	140 (38.3)
Aneurysm treatment - No treatment no. (%)	5 (1.4)	6 (1.6)
Median no. of days from first hospitalization to first aneurysm treatment (IQR)	1 (0-1)	1 (0-1)
Presence of delayed cerebral ischemia or radiographic vasospasm - Delayed cerebral ischemia no. (%)	50 (13.7)	57 (15.6)
Presence of delayed cerebral ischemia or radiographic vasospasm - Radiographic vasospasm no. (%)	134 (36.6)	128 (35.0)
Vasopressor or inotropic therapy no. (%)	180 (49.2)	155 (42.3)
Previous red-cell transfusion no. (%)	40 (10.9)	28 (7.7)
Hemoglobin level at randomization - g/dl Mean	9.4±0.6	9.3±0.6
Hemoglobin level at randomization - g/dl Median (IQR)	9.5 (9.1-9.8)	9.5 (9.0-9.8)

Arms

Field	Liberal Strategy Group	Control
Intervention	Mandatory red-cell transfusion at a hemoglobin level of ≤10 g per deciliter. The assigned transfusion strategy was applied from the time of randomization to 21 days after the first hospital presentation, death, or hospital discharge, whichever came first. Patients received leukoreduced red cells, 1 unit at a time, when the specified hemoglobin threshold was met.	Optional red-cell transfusion at a hemoglobin level of ≤8 g per deciliter. The assigned transfusion strategy was applied from the time of randomization to 21 days after the first hospital presentation, death, or hospital discharge, whichever came first. Patients received leukoreduced red cells, 1 unit at a time, when the specified hemoglobin threshold was met.
Duration	Up to 21 days of hospitalization, death, or hospital discharge	Up to 21 days of hospitalization, death, or hospital discharge

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
Unfavorable neurologic outcome at 12 months, defined as a score of 4 or higher on the modified Rankin scale (range, 0 to 6, with higher scores indicating greater disability).	Primary	136/361 (37.7%)	122/364 (33.5%)	0.88	0.22
12-month Functional Independence Measure (FIM) score (total score, all patients)	Secondary	79.8±54.5	82.8±54.6
12-month EuroQol five-dimension, five-level (EQ-5D-5L) utility index score (all patients)	Secondary	0.5±0.4	0.5±0.4
12-month Visual Analogue Scale (VAS) score (all patients)	Secondary	50±37.1	52.1±37.5
Mortality at 12 months	Secondary	98/361 (27.1%)	99/364 (27.2%)	0.99
New radiographic vasospasm	Secondary	149/366 (40.7%)	115/365 (31.5%)	0.77
New cerebral infarction	Secondary	65/366 (17.8%)	58/365 (15.9%)	0.89
New DCI (delayed cerebral ischemia)	Secondary	73/366 (19.9%)	64/365 (17.5%)	0.88
Acute transfusion reactions	Adverse	Similar incidence	Similar incidence
Prespecified adverse events	Adverse	Similar incidence	Similar incidence

Subgroup Analysis

Findings were consistent across prespecified subgroups for the primary outcome, including age (>55 years), sex, clinical severity (World Federation of Neurological Surgeons grade), radiographic grade (modified Fisher Scale), aneurysm securing method (neurosurgical vs. endovascular approach), and presence at baseline of delayed cerebral ischemia or radiographic vasospasm. Sensitivity analyses also showed consistent results.

Criticisms

Open-label design, meaning clinical teams were aware of the intervention, which could introduce bias (though minimized by blinded outcome assessment).
Ascertainment of some secondary outcomes (delayed cerebral ischemia, radiographic vasospasm, new cerebral infarction) is challenging in clinical trials due to overlapping definitions and incompletely understood pathophysiological features.
The trial may not have been powered to detect smaller, but clinically meaningful, differences in benefit or harm.
The study population, primarily from tertiary care centers in high-income countries, may limit generalizability to other settings.

Funding

Canadian Institutes of Health Research and others; Medical Research Futures Fund of Australia; Heart and Stroke Foundation (salary support to Dr. English).

Based on: SAHARA (N Engl J Med, 2025)

Authors: S.W. English, A. Delaney, D.A. Fergusson, ..., for the SAHARA Trial Investigators on behalf of the Canadian Critical Care Trials Group*

Citation: N Engl J Med 2025;392:1079-88. DOI: 10.1056/NEJMoa2410962

Content summarized and formatted by NeuroTrials.ai.

SAHARA

(2025)

Objective

Liberal (Hb ≤10 g/dL) versus restrictive (Hb ≤8 g/dL) transfusion strategies in patients with aneurysmal SAH and anemia.

Study Summary

• No significant difference between liberal and restrictive transfusion strategies in patients with SAH and anemia

Intervention

Liberal transfusion at Hb ≤10 g/dL vs. restrictive transfusion at Hb ≤8 g/dL for 21 days post admission.

Inclusion Criteria

Adults ≥18 years with aneurysmal SAH and Hb ≤10 g/dL within 10 days of admission. Excluded if unstable bleeding or nonaneurysmal SAH.

Study Design

Arms: Liberal Transfusion vs. Restrictive Transfusion

Patients per Arm: Liberal: 364, Restrictive: 361

Outcome

• Unfavorable outcome (mRS ≥4 at 12 months): 33.5% vs 37.7% (RR 0.88; 95% CI, 0.72–1.09; P=0.22). <br>• No significant differences in FIM, EQ-5D-5L, or adverse events..

Bottom Line

Major Points

The SAHARA trial was an investigator-initiated, pragmatic, open-label, randomized trial with a blinded outcome assessment.
A total of 742 patients were randomized at 23 centers across Canada, Australia, and the United States.
The primary outcome was an unfavorable neurologic outcome (modified Rankin scale score ≥4) at 12 months.
An unfavorable neurologic outcome occurred in 33.5% of the liberal-strategy group and 37.7% of the restrictive-strategy group (risk ratio, 0.88; 95% CI, 0.72 to 1.09; P=0.22).
Secondary outcomes, including functional independence (FIM score), quality of life (EQ-5D-5L utility index, VAS score), and mortality, did not show significant differences between groups.
Radiographic vasospasm was observed in 31.5% of the liberal-strategy group and 40.7% of the restrictive-strategy group (risk ratio, 0.77; 95% CI, 0.63 to 0.95).
The incidence of delayed cerebral ischemia and new cerebral infarction was similar in both groups.
Red-cell transfusions were received by 99.7% of the liberal-strategy group and 34.9% of the restrictive-strategy group.

Study Design

Study Type: Randomized Controlled Trial
Randomization: Yes
Blinding: Blinded outcome assessment by independent assessors who were unaware of trial-group assignments.
Sample Size: 742
Follow-up: 12 months
Centers: 23
Countries: Canada, Australia, United States

Primary Outcome

Definition: Unfavorable neurologic outcome at 12 months, defined as a score of 4 or higher on the modified Rankin scale (range, 0 to 6, with higher scores indicating greater disability).

Control	Intervention	HR/OR	P-value
136/361 (37.7%)	122/364 (33.5%)	0.88	0.22

Limitations & Criticisms

Open-label design, meaning clinical teams were aware of the intervention, which could introduce bias (though minimized by blinded outcome assessment).
Ascertainment of some secondary outcomes (delayed cerebral ischemia, radiographic vasospasm, new cerebral infarction) is challenging in clinical trials due to overlapping definitions and incompletely understood pathophysiological features.
The trial may not have been powered to detect smaller, but clinically meaningful, differences in benefit or harm.
The study population, primarily from tertiary care centers in high-income countries, may limit generalizability to other settings.

Citation

N Engl J Med 2025;392:1079-88. DOI: 10.1056/NEJMoa2410962

View Original Publication →

SAHARA

Liberal or Restrictive Transfusion Strategy in Aneurysmal Subarachnoid Hemorrhage

Clinical Question

Bottom Line

Major Points

Design

Inclusion Criteria

Exclusion Criteria

Baseline Characteristics

Arms

Outcomes

Subgroup Analysis

Criticisms

Funding

Ask about SAHARA