New England Journal of Medicine JUNE 20, 2013

Rapid Blood-Pressure Lowering in Patients with Acute Intracerebral Hemorrhage (INTERACT2)

Craig S. Anderson et al.

Bottom Line

In patients with acute intracerebral hemorrhage, intensive lowering of systolic blood pressure to less than 140 mm Hg did not significantly reduce the primary dichotomous outcome of death or major disability, though an ordinal analysis indicated a significant shift toward improved functional outcomes.

Key Findings

1. The primary outcome of death or major disability (modified Rankin scale 3-6) occurred in 52.0% of the intensive-treatment group compared to 55.6% of the standard-treatment group, which was not statistically significant (OR 0.87; 95% CI, 0.75 to 1.01; P=0.06).
2. An ordinal analysis of modified Rankin scale scores demonstrated a significant shift toward lower scores, indicating improved functional outcomes with intensive treatment (OR for greater disability, 0.87; 95% CI, 0.77 to 1.00; P=0.04).
3. Mortality at 90 days was virtually identical between the two groups (11.9% in the intensive group vs. 12.0% in the standard group).
4. Nonfatal serious adverse events were similar between the intensive (23.3%) and standard (23.6%) treatment groups, establishing the safety of rapid blood pressure lowering.

Study Design

Design
RCT
Open-Label
Sample
2,839
Patients
Duration
90 days
Median
Setting
21 countries
Population Adults with spontaneous intracerebral hemorrhage presenting within 6 hours of symptom onset with elevated systolic blood pressure (150 to 220 mm Hg) and no indication for immediate surgery.
Intervention Intensive blood pressure lowering targeting a systolic blood pressure of <140 mm Hg within 1 hour, using locally available intravenous antihypertensive agents.
Comparator Standard guideline-recommended blood pressure lowering targeting a systolic blood pressure of <180 mm Hg.
Outcome Death or major disability (defined as a modified Rankin scale score of 3 to 6) assessed at 90 days.

Study Limitations

The open-label administration of treatment (PROBE design) introduces potential performance bias, although outcome assessors were blinded.
The primary dichotomous endpoint barely missed statistical significance, making the reliance on the secondary ordinal analysis somewhat controversial.
A large proportion of the study population was enrolled in Asia (particularly China), which may limit the generalizability of the findings to populations with different baseline characteristics or stroke etiologies.
There was heterogeneity in the antihypertensive agents used, as medications were chosen at the discretion of the treating physicians rather than following a standardized algorithmic protocol.
Early crossover or failure to maintain targeted blood pressure parameters diluted the separation between the intervention and control arms.

Clinical Significance

INTERACT2 was a landmark trial that fundamentally shifted the management of acute intracerebral hemorrhage. It dispelled long-standing fears that rapidly lowering blood pressure would induce dangerous perihematomal ischemia. While it failed to meet its primary dichotomous endpoint of death or major disability, the clear safety profile and the favorable ordinal shift in functional independence influenced global stroke guidelines (AHA/ASA) to recommend targeting a systolic blood pressure of 140 mm Hg. This approach is now considered a safe and potentially effective standard of care to limit hematoma expansion and improve patient recovery.

Historical Context

For decades, the optimal blood pressure management in acute intracerebral hemorrhage was intensely debated. Elevated blood pressure is extremely common post-ICH and is associated with hematoma expansion and poor prognosis. However, many clinicians were hesitant to lower blood pressure aggressively out of fear that it would reduce cerebral perfusion pressure and exacerbate ischemic injury in the 'penumbra' surrounding the hematoma. Following the INTERACT1 pilot study which suggested safety and a reduction in hematoma growth, INTERACT2 was designed as the definitive Phase 3 trial. Its findings paved the way for subsequent trials like ATACH-2, which further refined target thresholds by showing that excessively aggressive lowering (to 110-139 mm Hg) provides no additional benefit and increases renal adverse events.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

What is the pathophysiological rationale for rapidly lowering blood pressure in acute intracerebral hemorrhage, and what historical physiological concern did the INTERACT2 trial help alleviate?

Key Response

Rapid blood pressure lowering aims to decrease hydrostatic pressure within the ruptured vessels, thereby preventing hematoma expansion, which is a major predictor of poor clinical outcomes. Historically, there was a theoretical concern that lowering blood pressure could reduce cerebral perfusion pressure and cause secondary ischemia in the 'perihematomal penumbra.' INTERACT2, along with other neuroimaging studies, demonstrated that rapid blood pressure reduction is safe and does not induce clinically significant perihematomal ischemia or worsen neurological outcomes.

Resident
Resident

Based on the INTERACT2 trial, if a patient presents with an acute spontaneous intracerebral hemorrhage and an initial systolic blood pressure of 195 mm Hg, what is your systolic blood pressure target, and what properties should your chosen intravenous antihypertensive possess?

Key Response

The target based on the INTERACT2 intervention is a systolic blood pressure of less than 140 mm Hg, achieved within 1 hour. To achieve this safely and avoid overshoot hypotension, the preferred medications are easily titratable, short-acting intravenous antihypertensives, such as nicardipine (a calcium channel blocker) or labetalol (a mixed alpha/beta blocker), which allow for smooth, continuous blood pressure control.

Fellow
Fellow

The INTERACT2 trial failed to show a statistically significant difference in its primary dichotomous outcome (mRS 3-6) but demonstrated a significant benefit in the prespecified ordinal analysis of the modified Rankin Scale. How does an ordinal shift analysis change the interpretation of neurologic recovery in intracerebral hemorrhage compared to a traditional dichotomous endpoint?

Key Response

A dichotomous endpoint lumps functional outcomes together (e.g., mRS 0-2 as 'good' vs 3-6 as 'bad'), which can miss clinically meaningful transitions, such as a patient shifting from mRS 5 (bedbound/incontinent) to mRS 4 (requiring assistance but able to walk). An ordinal analysis evaluates shifts across the entire spectrum of the modified Rankin Scale, providing greater statistical power and reflecting the clinical reality that any degree of functional improvement and preserved independence is highly valuable to ICH patients.

Attending
Attending

While INTERACT2 suggested a functional benefit for targeting a systolic blood pressure less than 140 mm Hg, the subsequent ATACH-2 trial was stopped early for futility and an increase in renal adverse events. How should an attending physician synthesize these two landmark trials when guiding the ICU team in acute intracerebral hemorrhage management?

Key Response

Synthesis requires understanding the nuances in trial execution: ATACH-2 achieved much more rapid, aggressive, and sustained blood pressure lowering (often driving SBP well below 130 mm Hg) using high-dose nicardipine, which led to a higher rate of acute kidney injury without improving neurological outcomes. The teaching point is that while aiming for an SBP of 140 mm Hg is reasonable and likely beneficial to prevent hematoma expansion (per INTERACT2), aggressive over-shooting below 130 mm Hg offers no added neuroprotection and actively increases the risk of systemic organ hypoperfusion.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

INTERACT2 utilized a proportional odds model for the ordinal analysis of the modified Rankin Scale, yielding a significant result despite a null primary dichotomous outcome. What is the critical assumption underlying the proportional odds model, and how might a violation of this assumption impact the validity of the trial's secondary conclusions?

Key Response

The critical assumption is the proportional odds (or parallel lines) assumption, which posits that the treatment effect (the odds ratio) is consistent across all possible cut-points of the ordinal scale. If this assumption is violated (e.g., the treatment significantly reduces mortality but paradoxically increases the number of patients in the severe disability category), relying on a single summary odds ratio becomes misleading, as it obscures heterogeneous treatment effects at different levels of stroke severity.

Journal Editor
Journal Editor

Given that INTERACT2 utilized a PROBE (Prospective Randomized Open, Blinded End-point) design, what specific threats to validity regarding performance bias would a peer reviewer scrutinize, particularly concerning the management of the standard-care control group?

Key Response

In an open-label trial of a critical care intervention, performance bias is a major concern. A reviewer would scrutinize whether the control group (SBP < 180 mm Hg) received less intensive overall ICU care, different co-interventions (such as less aggressive reversal of coagulopathy, different rates of intubation, or differing DVT prophylaxis), or earlier withdrawal of life-sustaining therapies compared to the intensively monitored intervention group. Ensuring that early Do-Not-Resuscitate orders and general medical management were balanced is essential to validate that the observed functional shift was solely due to blood pressure management.

Guideline Committee
Guideline Committee

How have the findings of INTERACT2, particularly when later contextualized by the ATACH-2 trial, influenced current AHA/ASA guidelines regarding the specific target and strength of recommendation for acute blood pressure lowering in spontaneous intracerebral hemorrhage?

Key Response

Based on INTERACT2 and ATACH-2, the 2022 AHA/ASA Guidelines for the Management of Patients With Spontaneous Intracerebral Hemorrhage state that for patients with mild-to-moderate ICH presenting with an SBP between 150 and 220 mm Hg, acute lowering of SBP to a target of 140 mm Hg is safe and can be effective for improving functional outcomes (Class 2a, Level of Evidence A). Crucially, referencing the harm seen in ATACH-2, the guidelines now also include a Class 3 (Harm) recommendation explicitly advising against lowering the SBP to less than 130 mm Hg due to the increased risk of renal injury.

Clinical Landscape

Noteworthy Related Trials

2008

INTERACT1 Trial

n = 404 · Lancet Neurol

Tested

Early intensive BP lowering (target systolic <140 mm Hg)

Population

Patients with acute spontaneous intracerebral hemorrhage

Comparator

Standard BP lowering (target systolic <180 mm Hg)

Endpoint

Proportional hematoma growth at 24 hours

Key result: Intensive BP lowering attenuated hematoma growth without increasing neurological deterioration or adverse events.
2008

FAST Trial

n = 841 · NEJM

Tested

Recombinant activated factor VII (rFVIIa)

Population

Patients with acute spontaneous intracerebral hemorrhage within 4 hours of onset

Comparator

Placebo

Endpoint

Death or severe disability at 90 days

Key result: Treatment with rFVIIa reduced hematoma growth but failed to improve clinical outcomes and increased arterial thromboembolic events.
2016

ATACH-2 Trial

n = 1,000 · NEJM

Tested

Intensive BP lowering (target systolic 110-139 mm Hg)

Population

Patients with acute intracerebral hemorrhage and hypertension

Comparator

Standard BP lowering (target systolic 140-179 mm Hg)

Endpoint

Death or disability (modified Rankin scale 4-6) at 3 months

Key result: Intensive BP lowering did not reduce death or disability compared to standard treatment and increased renal adverse events.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis