INTERACT2
Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial 2
Year of Publication: 2013
Authors: Anderson CS, Heeley E, Huang Y, et al.
Journal: New England Journal of Medicine
Citation: Anderson CS et al. N Engl J Med. 2013;368(25):2355-2365.
PDF: Clinical Question
Does intensive lowering of systolic blood pressure (SBP) to <140 mm Hg improve outcomes in patients with acute intracerebral hemorrhage (ICH)?
Study Overview
Objective
Determine whether early intensive blood pressure lowering improves outcomes in patients with acute spontaneous intracerebral hemorrhage (ICH).
Study Summary
In patients with spontaneous ICH and elevated systolic BP (150–220 mm Hg) within 6 hours of onset, intensive BP lowering to <140 mm Hg was safe and showed a modest improvement in functional outcomes, though it did not significantly reduce the primary composite outcome of death or major disability.
Intervention
International, multicenter, randomized, open-label trial with blinded outcome assessment. N=2839 patients randomized to intensive treatment (target SBP <140 mm Hg within 1 hour) vs. guideline-recommended treatment (target <180 mm Hg). Follow-up at 90 days.
Bottom Line
Intensive SBP lowering to <140 mm Hg within 6 hours of ICH onset was safe and associated with improved functional outcomes, though it did not significantly reduce death or major disability.
Major Points
- INTERACT2 was the first large randomized trial of acute BP management in intracerebral hemorrhage, establishing that targeting SBP <140 mmHg within 6 hours of ICH onset is SAFE and probably beneficial — transforming ICH management from permissive hypertension to early aggressive BP control.
- Primary outcome (death or major disability mRS ≥3 at 90 days): 52.0% intensive vs 55.6% standard (OR 0.87, 95% CI 0.75–1.01, p=0.06) — narrowly missed statistical significance by the conventional dichotomous analysis, but the trend was consistent and clinically meaningful.
- Pre-specified ordinal shift analysis of the full mRS distribution showed SIGNIFICANT benefit: common OR 0.87, p=0.04 — demonstrating that intensive BP lowering shifted outcomes favorably across the disability spectrum, not just at the death/disability threshold.
- 2,839 patients across 144 hospitals in 21 countries. Predominantly Asian (~68% Chinese) — the largest ICH trial at the time, though subsequently surpassed by INTERACT3 and ATACH-2.
- BP achieved: mean SBP 150 mmHg (intensive) vs 164 mmHg (standard) at 1 hour — a 14 mmHg difference. Intensive arm reached target <140 mmHg within 1 hour in 33.4% of patients, highlighting the practical difficulty of achieving rapid BP control.
- No increase in adverse events: nonfatal SAEs were 23.3% (intensive) vs 23.6% (standard) (p=0.92) — directly addressing the longstanding fear that aggressive BP lowering would cause perihematomal ischemia or renal failure.
- Hematoma growth: absolute hematoma growth was numerically smaller in the intensive group, though not statistically significant (between-group adjusted absolute difference 1.4 mL, P=0.18; relative difference 4.5%, P=0.27) — mechanistically supporting the theory that early BP control limits hemorrhage expansion.
- Median time to randomization was 3.7 hours — leaving a window where some hematoma expansion may have already occurred. Earlier BP lowering (within 1–2 hours) might show greater benefit, as tested in subsequent trials.
- Set the standard of care: AHA/ASA 2015 ICH guidelines gave intensive BP lowering to SBP <140 mmHg a Class IIa recommendation based primarily on INTERACT2, making it the de facto approach in most stroke centers worldwide.
- Followed by ATACH-2 (2016) which tested SBP <140 vs <180 using IV nicardipine — found NO benefit but MORE renal events, suggesting there may be a 'floor effect' where additional BP lowering beyond a certain point is harmful. INTERACT2's less aggressive approach (targeting <140 within 1 hour) appears to be the sweet spot.
Design
Study Type: Open-label, randomized, controlled trial with blinded endpoint assessment
Randomization: 1
Blinding: Endpoint-assessor blinded
Enrollment Period: 2008 – 2012
Follow-up Duration: 90 days
Centers: 144
Countries: Australia, China, UK, Vietnam, India, South Korea, Pakistan, Chile, Mexico, Brazil, and others
Sample Size: 2839
Analysis: Modified intention-to-treat
Inclusion Criteria
- Adults ≥18 years with spontaneous intracerebral hemorrhage confirmed by CT scan within 6 hours of symptom onset.
- Elevated systolic blood pressure: SBP 150–220 mmHg (patients had to have elevated BP but not extreme hypertension requiring immediate IV therapy).
- Clinical uncertainty about the benefit of intensive BP lowering — treating physician felt equipoise about the optimal BP target (uncertainty principle).
- Randomization possible within 6 hours of stroke onset — fast enrollment was essential to test the 'early BP lowering prevents hematoma expansion' hypothesis.
Exclusion Criteria
- Secondary ICH causes: ruptured aneurysm, arteriovenous malformation, tumor, trauma, hemorrhagic transformation of ischemic stroke, or dural venous sinus thrombosis.
- GCS score ≤5 (very deep coma) — these patients had very poor prognosis regardless of BP management.
- Planned early surgical hematoma evacuation — surgery would confound the BP intervention effect.
- Massive hematoma with imminent brain herniation — immediate surgical/medical management took priority.
- SBP >220 mmHg at presentation — these patients were thought to require immediate IV antihypertensive regardless of randomization.
- SBP <150 mmHg at presentation — below the inclusion threshold; no intervention needed.
- Clear indication for anticoagulation that could not be interrupted — complicating hematoma management.
- Known terminal illness with life expectancy <6 months.
- Previous participation in INTERACT2 or concurrent enrollment in another ICH intervention trial.
Baseline Characteristics
| Characteristic | Control | Active |
|---|---|---|
| Age (mean) | 64.1 | 63.0 |
| Male (%) | 61.7% | 64.2% |
| Time to randomization (median) | 3.7 h | 3.7 h |
| Baseline SBP (mean) | 179 mm Hg | 179 mm Hg |
| GCS score (median) | 14 | 14 |
| Volume of hematoma (median) | 11 mL | 11 mL |
| Intraventricular extension (%) | 28.0% | 28.7% |
| Location (deep) | 83.2% | 83.8% |
Arms
| Field | Control | Intensive BP Lowering (target SBP <140 mmHg within 1 hour) |
|---|---|---|
| Intervention | Target SBP <180 mmHg per 2003 AHA/ASA ICH guidelines. Antihypertensive agents chosen by treating physician per local protocols. Typically IV labetalol, nicardipine, or hydralazine for acute BP reduction; oral agents transitioned as tolerated. Achieved mean SBP of 164 mmHg at 1 hour — substantially below the 180 target, reflecting real-world tendency to treat BP even in the 'standard' arm. | Target SBP <140 mmHg within 1 hour of randomization and maintained for ≥7 days. Protocol-flexible: any IV or oral antihypertensive agent permitted, chosen by treating physician. Most common agents were IV urapidil (in Asia), IV labetalol, IV nicardipine, and oral amlodipine. Achieved mean SBP of 150 mmHg at 1 hour (33.4% reached target <140 mmHg). The flexible agent approach contrasted with ATACH-2's mandated IV nicardipine protocol. |
| Duration | Acute phase up to 7 days, with 90-day outcome assessment | Acute phase up to 7 days, with 90-day outcome assessment |
Outcomes
| Outcome | Type | Control | Intervention | HR / OR / RR | P-value |
|---|---|---|---|---|---|
| Death or major disability (modified Rankin score ≥3) at 90 days | Primary | 55.6% | 52.0% | 0.87 | 0.06 |
| Ordinal mRS distribution | Secondary | Worse shift | Favorable shift | OR 0.87 (0.77–1.00) | 0.04 |
| Death at 90 days | Secondary | 12.0% | 11.9% | OR 0.99 (0.79–1.25) | 0.96 |
| Any serious adverse event | Adverse | 23.6% | 23.3% | 0.92 | |
| Renal adverse event | Adverse | 0.3% | 0.6% | 0.12 |
Criticisms
- Primary endpoint narrowly missed significance (p=0.06) by dichotomous analysis — this 'near-miss' left room for skeptics to argue the benefit was not proven. However, the pre-specified ordinal analysis was significant (p=0.04), and most subsequent guidelines accepted the totality of evidence.
- Open-label design — physicians knew the BP target, potentially influencing overall care intensity and outcome assessment. PROBE design (blinded endpoint assessment) mitigated this, but treatment contamination (standard arm patients receiving more aggressive BP treatment than protocol) was documented.
- ~68% Chinese enrollment — may limit generalizability. Chinese patients may have different ICH etiology (more hypertensive vasculopathy), response to antihypertensives, and background care practices than Western populations.
- Moderate BP difference achieved: 14 mmHg at 1 hour (150 vs 164 mmHg) — the standard arm was treated more aggressively than protocol required (<180 target), narrowing the between-group difference and potentially diluting the treatment effect.
- No specific antihypertensive protocol — flexible agent choice improves generalizability but introduces heterogeneity. Different agents (nicardipine vs labetalol vs urapidil) have different hemodynamic profiles and may affect perihematomal perfusion differently.
- 90-day follow-up only — long-term functional recovery after ICH often continues beyond 3 months. Longer follow-up might have shown different results.
- Did not assess hematoma expansion by serial imaging in all patients — the mechanistic link between BP lowering and reduced expansion was inferred but not directly proven in this trial.
- Excluded GCS ≤5 patients — the most severe ICH patients, who have the highest mortality and might theoretically benefit most (or least) from BP control, were not studied.
- Subsequently challenged by ATACH-2 (2016): same BP target (<140 mmHg) but with mandatory IV nicardipine and faster BP reduction — showed NO benefit and MORE renal events, raising questions about whether speed/method of BP lowering matters as much as the target itself.
Funding
National Health and Medical Research Council of Australia, Chinese Ministry of Health, and multiple international funding agencies
Based on: INTERACT2 (New England Journal of Medicine, 2013)
Authors: Anderson CS, Heeley E, Huang Y, et al.
Citation: Anderson CS et al. N Engl J Med. 2013;368(25):2355-2365.
Content summarized and formatted by NeuroTrials.ai.