← Back

STOP II

Discontinuing Prophylactic Transfusions Used to Prevent Stroke in Sickle Cell Disease

Share Share Copied! Compare

Year of Publication: 2005

Authors: The Optimizing Primary Stroke Prevention in Sickle Cell Anemia (STOP 2) Trial Investigators

Journal: The New England Journal of Medicine

Citation: N Engl J Med 2005;353:2769-78.

Link: https://doi.org/10.1056/NEJMoa050460

PDF: https://www.nejm.org/doi/pdf/10.1056/NEJMoa050460

Clinical Question

In children with sickle cell anemia at high risk for stroke whose transcranial Doppler (TCD) velocities have normalized after at least 30 months of chronic prophylactic transfusions, can transfusions be safely discontinued?

Study Overview

Objective

Determine whether chronic transfusion therapy can be safely discontinued in children with sickle cell anemia and initially abnormal TCD velocities that normalize after transfusion.

Study Summary

In children with sickle cell disease and previously abnormal TCD velocities that normalized after transfusion, discontinuing transfusion was associated with a high risk of reversion to abnormal velocities and recurrent stroke. Continued transfusion remained effective at preventing stroke.

Intervention

Randomized, multicenter trial of 79 children with sickle cell disease and abnormal baseline TCD who normalized after ≥30 months of transfusions. Randomized to continue transfusion vs. discontinue and observe. TCD monitoring and stroke surveillance were ongoing.

Bottom Line

Discontinuing prophylactic transfusions in children with sickle cell anemia, even after their TCD velocities have normalized, results in a high rate of reversion to abnormal high-risk velocities and stroke. The trial was stopped early for safety reasons, indicating that transfusion therapy should be continued indefinitely in this high-risk population.

        Major Points
        STOP II definitively answered a critical clinical question: Can you STOP chronic transfusions after TCD normalizes in sickle cell disease? The answer is NO — 39% (16/41) reached the composite endpoint within a mean of only 4.5 months among those who had events.
Trial stopped prematurely at the fourth interim analysis after 79 of planned 100 patients — 16/41 (39%) in the transfusion-halted group had events (14 abnormal TCD, 2 strokes) vs 0/38 in continued transfusion (P<0.001).
Companion to the original STOP trial (NEJM 1998), which showed TCD-guided prophylactic transfusion reduces stroke incidence in sickle cell children from ~10%/year to <1%/year. STOP II proves this benefit requires INDEFINITE transfusion.
Among event patients, mean time to primary endpoint was 4.5±2.6 months (median 3.2 months, range 2.1–10.1). Suggests ongoing vascular injury persists despite TCD normalization, and that TCD improvement reflects reduced blood velocity (lower HbS%) rather than structural vascular healing.
The only baseline predictor of events was the pre-transfusion TCD velocity (average of the two screening velocities before transfusion; P=0.05) — patients with higher initial velocities were at greater risk of reversion, suggesting a dose-response relationship between vascular disease severity and treatment dependence.
Established the clinical paradigm that chronic transfusion for stroke prevention in SCD is a LIFELONG commitment — raising major concerns about iron overload, alloimmunization, and transfusion access disparities.
Led directly to investigation of alternatives to chronic transfusion: TWiTCH (NEJM 2016) showed hydroxyurea could replace transfusion in selected children.
Iron overload was substantial in both groups at baseline (mean ferritin ~3000–3300 ng/ml); by 12 months, continued-transfusion ferritin rose to 3562±1536 vs halted 1832±916 ng/ml (P=0.002), highlighting the urgent need for chelation and alternative stroke prevention strategies.
All patients had received ≥30 months of prophylactic transfusions (≥24 transfusions in 30 months and HbS <30% in at least 20 of the 30 months) — the finding that this substantial treatment duration was insufficient for 'cure' was surprising and clinically important.
Exclusion of patients with moderate-to-severe MRA lesions means STOP II applies to those with functional (hemodynamic) rather than structural vascular disease. Patients with fixed stenoses likely need even more aggressive management.

      

Design

Study Type: Randomized, controlled trial.

Randomization: 1

Blinding: Blinded adjudication of TCD and imaging endpoints.

Enrollment Period: Not specified in the paper; trial was stopped early by the NHLBI/DSMB at the fourth interim analysis due to safety concerns.

Follow-up Duration: Planned as a 54-month study with 18 months of follow-up after recruitment ended. Trial stopped early at fourth interim analysis. Among the 16 transfusion-halted patients who had a primary endpoint, mean time from randomization to event was 4.5±2.6 months (median 3.2, range 2.1–10.1); 16/41 halted patients remained event-free at study end, 8 of whom were followed for >25 months.

Centers: Multiple centers involved in the original STOP trial, including U.S. and Canadian sites.

Countries: United States, Canada

Sample Size: 79

Analysis: Event rates were compared with a log-rank test, with data censored at crossover (initiation of hydroxyurea or resumption of regular transfusion in the halted group) — this is not a pure intention-to-treat analysis. Baseline characteristics compared using Student's t-test, chi-square, or Fisher's exact test; laboratory values at 6 and 12 months compared with baseline by Student's t-test. All P values two-sided and not adjusted for multiple testing. Interim analyses used a Lan-DeMets spending function approximating an O'Brien–Fleming boundary.

Inclusion Criteria

Children with sickle cell disease (HbSS or HbSβ0) at high risk for stroke based on prior abnormal TCD.
Received prophylactic blood transfusions for at least 30 months (≥24 transfusions in 30 months and HbS <30% in ≥20 of the 30 months).
Two normal TCD examinations at least 2 weeks apart while receiving transfusions within 4 months of randomization.
Age 5–20 years at randomization.
Written informed consent from a parent/guardian; child assent where appropriate.

Exclusion Criteria

Prior stroke.
Indication for chronic transfusion for reasons other than stroke prevention.
Contraindication to chronic transfusion.
Moderate-to-severe intracranial arterial disease on MRA.

Baseline Characteristics

Characteristic	Control	Active
Age-yr (mean)	12.5 ± 3.3	12.0 ± 3.1
Male sex no. (%)	20 (53)	13 (32)
Lesions on initial MRI - no. of patients (%)	10 (26)	11 (27)
Systolic blood pressure - mm Hg	113 ± 12	109 ± 12
Hemoglobin S - %	21.0 ± 8.6	19 ± 11
Ferritin - ng/ml	3274 ± 1718	3005 ± 1504

Arms

Field	Control	Transfusion Halted
Intervention	Continued regimen of prophylactic blood transfusions to maintain hemoglobin S levels below 30%.	Discontinuation of prophylactic blood transfusions, with continued TCD monitoring and clinical surveillance. Transfusions could be reinitiated if TCD velocities became abnormal.
Duration	Ongoing until study end (32/38 still receiving transfusions at end of trial)	Variable — 16/41 reached primary endpoint (mean 4.5±2.6 months); 9/41 censored at crossover to transfusion or hydroxyurea; 16/41 remained event-free at study end, including 8 followed >25 months.

Outcomes

Outcome	Type	Control	Intervention	P-value
A composite of stroke (cerebral infarction or intracranial hemorrhage) or reversion to abnormal velocity on transcranial Doppler ultrasonography (defined as two consecutive studies with abnormal velocities, three consecutive studies with average velocity ≥200 cm/sec, or three consecutive inadequate studies plus severe stenosis on MRA).	Primary	0% (0/38)	39% (16/41)	<0.001
	Secondary	9.4 ± 0.9 (n=31)	7.7 ± 0.8 (n=23)	<0.001
	Secondary	27.8 ± 2.5	22.0 ± 2.8	<0.001
	Secondary	8.9 ± 3.3	14.8 ± 5.7	<0.001
	Secondary	25.4 ± 10.9	81.0 ± 13.6	<0.001
	Secondary	2.6 ± 2.1	7.0 ± 4.3	<0.001
	Secondary	3555 ± 2043	2604 ± 1394	0.06
	Secondary	3562 ± 1536 (n=25)	1832 ± 916 (n=11)	0.002
	Secondary	469 ± 164	616 ± 240	0.046
	Secondary	2.3 ± 1.6	4.6 ± 3.3	0.058
Reversion to abnormal TCD velocity	Adverse	0% (0/38)	34% (14/41)
Stroke	Adverse	0% (0/38)	5% (2/41)
Death (complications of acute chest syndrome)	Adverse	1/38	0/41
Transfusion reactions (in 7 patients; 1 serious requiring hospitalization)	Adverse	9 reactions total across the study	9 reactions total across the study
New alloimmunization (anti-Kpa)	Adverse	1/38	0/41
Hepatitis C seroconversion	Adverse	0	0	0/68 tested at end of study
Receiving chelation at end of trial	Adverse	35/38 (93%)	31/41 (76%)
Acute chest syndrome after transfusion stopped (halted group)	Adverse	Not reported	18/41 had ≥1 episode	0.22 (not a predictor of primary endpoint)

Criticisms

Early termination (79 of 100 planned) — while ethically essential, limits precision of effect estimates and prevents assessment of longer-term outcomes in the continued-transfusion group.
Small sample size limits power for subgroup analyses — cannot identify which patients (if any) might safely discontinue transfusion.
Does not address alternatives to chronic transfusion — the question of whether hydroxyurea could substitute was unanswered until TWiTCH (2016). STOP II only compared continuing vs stopping transfusion.
The primary endpoint combined TCD reversion with stroke — most events were TCD reversion (14/16), not stroke (2/16). TCD reversion is a surrogate that may overestimate clinical risk.
Analysis censored at crossover (hydroxyurea or resumed transfusion in the halted group), so this is not a pure intention-to-treat analysis and may underestimate the risk of transfusion cessation.
No long-term follow-up beyond the short study period — cannot determine if continued transfusion patients eventually develop complications (iron overload, alloimmunization) that offset stroke prevention benefit.
Iron overload was already severe at enrollment (mean ferritin ~3000–3300 ng/mL) — the study did not systematically address chelation or assess organ damage from iron burden.
Excluded patients with moderate-to-severe MRA lesions — these highest-risk patients may have different treatment dynamics. Results may not apply to children with structural vasculopathy.
Generalizability limited to HbSS/HbSβ0 sickle cell anemia — does not address HbSC or milder sickle-beta thalassemia, which have different stroke risk profiles.
The implication of indefinite transfusion raises equity concerns — chronic transfusion requires reliable blood supply, insurance coverage, and regular medical access, creating disparities in low-resource settings.

Funding

National Heart, Lung, and Blood Institute (NHLBI) grants U01 HL 052193 and U01 HL 052016.

Based on: STOP II (The New England Journal of Medicine, 2005)