← Back

SAINT II

NXY-059 for the Treatment of Acute Ischemic Stroke

Share Share Copied! Compare

Year of Publication: 2007

Authors: Ashfaq Shuaib, Kennedy R. Lees, Patrick Lyden, ..., for the SAINT II Trial Investigators

Journal: New England Journal of Medicine

Citation: N Engl J Med 2007;357:562-71

Link: https://doi.org/10.1056/NEJMoa070240

PDF: https://www.nejm.org/doi/pdf/10.1056/NEJ...ticleTools=true

Clinical Question

Does the free-radical–trapping neuroprotectant NXY-059, given within 6 hours of acute ischemic stroke, reduce disability at 90 days compared with placebo?

Study Overview

Objective

To confirm whether the free-radical–trapping neuroprotectant NXY-059, given within 6 hours of acute ischemic stroke onset, reduces disability at 90 days, following a promising signal from the SAINT I trial.

Study Summary

NXY-059 did not reduce disability: mRS distribution at 90 days showed no significant difference vs placebo (P = 0.33; OR 0.94, 95% CI 0.83–1.06)
Trichotomized mRS analysis confirmed no benefit (OR 0.92, 95% CI 0.80–1.06)
No benefit on any secondary end point (NIHSS, Barthel index at 90 days)
No reduction in alteplase-related symptomatic or asymptomatic intracranial hemorrhage
Mortality was equal in both groups (267 deaths each arm); adverse event rates were similar
NXY-059 is ineffective for the treatment of acute ischemic stroke

Intervention

IV NXY-059 72-hour infusion (initial rate 2270 mg/hr then 480–960 mg/hr for 71 hours, adjusted for creatinine clearance, targeting plasma unbound concentration of 260 µmol/L) vs matching placebo, started within 6 hours of stroke onset

Patients per Arm

NXY-059: 1588; Placebo: 1607 (efficacy analysis); 3306 total randomized

Bottom Line

NXY-059 is definitively ineffective for the treatment of acute ischemic stroke within 6 hours of symptom onset; it should not be used as a neuroprotective agent in this setting, and its null result underscores the persistent translational gap between animal model neuroprotection and human clinical benefit.

        Major Points
        NXY-059 did not reduce disability: mRS distribution at 90 days showed no significant difference vs placebo (P = 0.33; OR for limiting disability 0.94, 95% CI 0.83–1.06).
Trichotomized mRS analysis (0–1 vs 2–3 vs 4–5, with deaths assigned a score of 5) confirmed no benefit (OR 0.92, 95% CI 0.80–1.06).
No benefit was observed on any secondary end point, including NIHSS and Barthel index at 90 days.
Among the ~44% of patients receiving concomitant alteplase, NXY-059 did not reduce the frequency of symptomatic or asymptomatic intracranial hemorrhage.
Mortality was equal in both groups (267 deaths per arm); adverse event rates were similar (84.2% NXY-059 vs 84.5% placebo).
Target plasma concentrations of NXY-059 were achieved in 96.6% of treated patients, ruling out pharmacokinetic failure as an explanation for the null result.
SAINT II (n=3306, 362 centers, 31 countries) was robustly powered and confirmed that the positive SAINT I signal was not reproducible.
This trial is a landmark example of the failure to translate neuroprotective efficacy from animal stroke models to human clinical benefit.

      

Design

Study Type: Randomized, double-blind, placebo-controlled trial

Randomization: 1

Blinding: Double-blind

Allocation: Computer-generated coding system; stratified by country, baseline NIHSS score, side of infarction, and intention to treat with alteplase

Enrollment Period: May 2003 – June 2006

Follow-up Duration: 90 days

Centers: 362

Countries: Argentina, Australia, Austria, Belgium, Brazil, Bulgaria, Canada, Chile, China, Czech Republic, France, Germany, Greece, Hong Kong/China, Hungary, Israel, Korea, Mexico, Philippines, Poland, Portugal, Russia, Singapore, Slovakia, South Africa, Spain, Switzerland, Taiwan, United Kingdom, United States

Sample Size: 3306

Analyzed: 3195

Analysis: Modified intent-to-treat (all patients who commenced any treatment and had any post-treatment assessment data); Cochran–Mantel–Haenszel test with modified ridit scores adjusted for stratification variables; hierarchical secondary outcome testing

Power Calculation: Sample size of 3200 patients provided ≥80% power to detect a common OR of 1.2 across all mRS cutoff points (the effect seen in SAINT I)

Registration: NCT00061022

Inclusion Criteria

Age ≥18 years
Clinical diagnosis of acute ischemic stroke with onset within the previous 6 hours
NIHSS score ≥6 with at least 2 points attributable to limb weakness
Informed consent from patient or acceptable surrogate

Exclusion Criteria

Creatinine clearance <30 ml per minute (treatment withdrawn if identified post-enrollment)
For patients receiving alteplase: inability to start study drug within 30 minutes after completion of alteplase infusion

Baseline Characteristics

Characteristic	Placebo (n=1631)	NXY-059 (n=1610)
Mean age (yr)	69	68.8
Male sex	53.3% (870/1631)	55.8% (898/1610)
Mean time from onset to treatment (hr:min)	3:49	3:46
Mean NIHSS score	13	13
Alteplase treatment	43.9% (716/1631)	44.1% (710/1610)
Mean time from onset to alteplase (hr:min)	2:24	2:22
Mean age in alteplase subgroup (yr)	68.6	67.7
Mean NIHSS in alteplase subgroup	13.8	14.1
Hypertension	76.4% (1246/1631)	77.2% (1243/1610)
Prior stroke	21.8% (356/1631)	22.2% (357/1610)
Prior TIA	9.9% (161/1631)	8.4% (135/1610)
Ischemic heart disease	32.9% (536/1631)	33.0% (531/1610)
Atrial fibrillation	30.6% (499/1631)	27.2% (438/1610)
Diabetes mellitus	24.9% (406/1631)	24.0% (386/1610)
Antiplatelet drug use	34.7% (566/1631)	36.6% (590/1610)
Cardioembolic stroke	47.6% (775/1627; denominator per table footnote)	46.1% (738/1600; denominator per table footnote)

Arms

Field	Control	NXY-059
N	1607	1588
Intervention	Matching placebo IV infusion for 72 hours, started within 6 hours of stroke onset	IV NXY-059 72-hour infusion: initial rate 2270 mg/hr, reduced after 1 hour to 480–960 mg/hr (32–64 ml/hr) for remaining 71 hours, adjusted for creatinine clearance, targeting unbound plasma concentration of 260 µmol/L
Duration	72 hours	72 hours

Outcomes

Outcome	Type	Control	Intervention	HR / OR / RR	P-value
Distribution of scores on the modified Rankin scale (mRS, range 0–5 with deceased patients scored 5) at 90 days, analyzed across the whole distribution using the Cochran–Mantel–Haenszel test with modified ridit scores, adjusted for stratification variables	Primary	Placebo (n=1607): 1313 assessed at day 90, 267 died (mRS assigned 5), 27 lost to follow-up with last observation carried forward	NXY-059 (n=1588): 1279 assessed at day 90, 267 died (mRS assigned 5), 42 lost to follow-up with last observation carried forward	0.94	0.33
Trichotomized mRS at 90 days (scores 0–1 vs 2–3 vs 4–5; deaths assigned a score of 5)	Secondary			0.92
NIHSS score at 90 days (first prespecified secondary end point), analyzed by Cochran–Mantel–Haenszel test with modified ridit score adjusted for baseline NIHSS, side of infarct, and alteplase use	Secondary	No evidence of efficacy (P = 0.73; Mann–Whitney U 0.50, 95% CI 0.48–0.52); formal testing not reached in hierarchy because primary end point was not significant
Barthel index ≥95 at 90 days (activities of daily living, 0–100)	Secondary	No evidence of efficacy (40.9% NXY-059 vs 42.3% placebo)
mRS at 7 and 30 days	Secondary	No evidence of efficacy
Symptomatic intracranial hemorrhage in alteplase-treated patients (worsening NIHSS ≥4 points within 36 hours plus blood on imaging)	Secondary	No difference (4.6% NXY-059 vs 5.3% placebo; P = 0.57)
Asymptomatic intracranial hemorrhage in alteplase-treated patients	Secondary	No difference (17.9% NXY-059 vs 16.1% placebo, post hoc analysis); overall cerebral hemorrhage 22.5% vs 21.4% (P = 0.60)
Mortality at 90 days	Safety	Equal in both groups: 267 deaths (16.5%) per arm; mean time to death 24.3±1.4 days (NXY-059) vs 21.9±1.4 days (placebo), P = 0.98 by log-rank
Overall adverse events	Safety	Similar between groups: 84.2% NXY-059 vs 84.5% placebo
Serious adverse events	Safety	Similar between groups: 39.6% NXY-059 vs 40.2% placebo
Study drug discontinuation	Safety	4.5% NXY-059 vs 6.3% placebo
Denominators	Adverse	Safety population — Placebo N=1631, NXY-059 N=1610
Overall Rates	Adverse	Any AE: 84.2% NXY-059 vs 84.5% placebo; SAE: 39.6% vs 40.2%. Only hypokalemia occurred significantly more often with NXY-059; it was mild, asymptomatic, and resolved within 7 days.
Most Common Serious Adverse Events (Placebo vs NXY-059)	Adverse	Stroke in evolution: 9.6% (157) vs 9.3% (150) · Ischemic stroke: 3.8% (62) vs 3.4% (55) · Pneumonia: 2.3% (37) vs 2.4% (39) · Hemorrhagic transformation of stroke: 1.7% (28) vs 1.9% (30) · Aspiration pneumonia: 1.2% (20) vs 1.6% (26) · Brain edema: 1.7% (28) vs 1.5% (24) · Cerebral hemorrhage: 1.5% (25) vs 1.4% (23) · Sepsis: 0.6% (9) vs 1.4% (22) · Myocardial infarction: 1.5% (25) vs 1.1% (18) · Atrial fibrillation: 0.4% (7) vs 1.0% (16)
Most Common Adverse Events (Placebo vs NXY-059)	Adverse	Pyrexia: 16.4% (268) vs 17.9% (289) · Headache: 13.9% (227) vs 15.9% (256) · Hypokalemia: 9.2% (150) vs 12.1% (195) — only AE significantly higher with NXY-059 · Constipation: 9.3% (151) vs 10.6% (170) · Urinary tract infection: 9.9% (161) vs 10.2% (164) · Stroke in evolution: 10.1% (164) vs 9.4% (152) · Atrial fibrillation: 6.3% (103) vs 6.9% (111) · Nausea: 6.9% (113) vs 6.9% (111) · Pneumonia: 5.5% (90) vs 6.4% (103) · Insomnia: 4.8% (79) vs 6.3% (101)
Note	Adverse	Table 2 lists the 10 most common SAEs and 10 most common AEs in the NXY-059 group. Formal statistical testing was not conducted given >100 potential events; a between-group difference of ≥1.5% approximates P = 0.05 before multiplicity adjustment.

Subgroup Analysis

Prospectively planned analysis of intracranial hemorrhage in the ~44% of patients receiving concomitant alteplase showed no reduction in symptomatic or asymptomatic hemorrhage with NXY-059 vs placebo, contradicting the post hoc SAINT I signal. No significant treatment interaction with time to treatment, alteplase use, diabetes, hypertension, or stroke severity.

Criticisms

The positive SAINT I result was not reproducible in a much larger, better-powered replication — raising the possibility that the original finding was a false positive driven by smaller sample size.
Despite robust pharmacokinetic confirmation (96.6% of patients achieving ≥150 µmol/L, well above neuroprotective animal model thresholds), no clinical benefit emerged, indicating the drug-target mechanism itself may not translate to human stroke.
The trial highlights the ongoing translational failure in stroke neuroprotection: dozens of agents that worked in animal models have failed in human trials, pointing to fundamental differences in the models.
Enrollment spanned 31 countries and 362 centers, introducing heterogeneity in care practices; however, prognostic factors were well balanced and the result was uniformly null.

Funding

AstraZeneca (sponsor; responsible for operational aspects including data collection, storage, and analysis per approved plan). NXY-059 is subject to a partnership agreement between AstraZeneca and Renovis; Renovis had no influence on trial conduct, analysis, or interpretation but was given the opportunity to review and comment on the manuscript.

Based on: SAINT II (New England Journal of Medicine, 2007)