The Clotbust Review

Impact of Ultra-Early Perioperative Antihypertensive Therapy in Acute Intracerebral Hemorrhage — INTERACT3 Secondary Analysis

· DOI: 10.1161/STROKEAHA.125.053989 · PMC13196860 · stroke deep-dive intracerebral-hemorrhage blood-pressure perioperative INTERACT3

Stylized illustration of ultra-early blood pressure control in acute intracerebral hemorrhage
AI-generated editorial illustration. Generated for editorial use.

Clinical Question (PICO)

In patients with acute spontaneous intracerebral hemorrhage (ICH) who undergo surgical hematoma evacuation, does achieving intensive systolic blood pressure (SBP) targets (<140 mm Hg) within 2 hours of treatment initiation — compared with delayed control (>2 hours) — reduce 6-month mortality, death or disability, and serious adverse events? (Post-hoc secondary analysis of the INTERACT3 cluster-randomized trial; not a pre-specified subgroup.)

Bottom Line

Among 1506 INTERACT3 participants who underwent surgical hematoma evacuation (mean age 59.7 years, 33.9% women), reaching SBP <140 mm Hg within 2 hours of arrival was not associated with a statistically significant reduction in 6-month mortality (adjusted HR 0.81 [95% CI 0.63–1.04]; P=0.09) but was associated with lower death or disability at 6 months (adjusted OR 0.71 [95% CI 0.56–0.90]; P=0.01), a favorable mRS shift (adjusted common OR 0.73 [95% CI 0.60–0.89]; P<0.01), better EuroQol-5D-3L mobility/pain/usual-activities scores, and fewer serious adverse events (adjusted OR 0.73 [95% CI 0.57–0.94]; P=0.01). The authors conclude perioperative ultra-early intensive BP reduction is feasible and safe, with time-sensitive benefit on functional and quality-of-life outcomes.

Design

  • Trial type: Post-hoc secondary analysis of a pragmatic, international, multicenter, blinded-endpoint, stepped-wedge cluster-randomized trial (INTERACT3). Reported per STROBE checklist.
  • N (surgical sub-cohort): 1506 patients who underwent hematoma evacuation, drawn from the 7036 INTERACT3-enrolled ICH patients.
    • ≤2 h to SBP <140 mm Hg: subgroup analyzed

    • 2 h to SBP <140 mm Hg: subgroup analyzed

  • Exposure (not randomized): Time from hospital arrival to first achieving SBP <140 mm Hg (≤2 h vs >2 h).
  • Setting: 121 hospitals in 9 low- and middle-income countries (Brazil, China, India, Mexico, Nigeria, Pakistan, Peru, Sri Lanka, Vietnam) and Chile.
  • Enrollment: 2017–2021 (INTERACT3 main trial window).
  • Analysis: Cox regression (mortality), logistic regression (binary outcomes), ordinal logistic regression (mRS shift), linear regression (continuous HRQoL scores). Adjusted for age, sex, treatment type, and admission Glasgow Coma Scale score.
  • Primary outcome: 6-month all-cause mortality.

Population

Inclusion Criteria (INTERACT3; this analysis)

  • Adults with acute spontaneous intracerebral hemorrhage confirmed on imaging
  • Within hours of symptom onset
  • Underwent surgical hematoma evacuation as part of routine care at the enrolling hospital

Exclusion Criteria

  • ICH secondary to trauma, aneurysm, vascular malformation, tumor, or coagulopathy reversal
  • Patients managed non-operatively (excluded from this surgical sub-analysis only; they remain in the parent trial)

Baseline Characteristics (surgical sub-cohort, n=1506)

  • Mean (SD) age: 59.7 (11.8) years
  • Women: 33.9%
  • Country mix dominated by China, India, Pakistan, Vietnam (per INTERACT3 geography); the parent trial was 65.3% Asian, 21.3% Latin American, 13.4% other.
  • Admission GCS and hematoma volume distribution not reported in the abstract; baseline covariates adjusted for in the models are age, sex, treatment type, and admission GCS.

Interventions

  • ≤2 h group: Intensive BP control achieving SBP <140 mm Hg within 2 hours of hospital arrival (intravenous nicardipine, labetalol, urapidil, glyceryl trinitrate, or local agent per site).
  • >2 h group: Same target (SBP <140 mm Hg) achieved after 2 hours of treatment initiation.
  • Both arms received the full INTERACT3 care bundle: glucose control (target 6.1–7.8 mmol/L unless known diabetes), temperature control (<37.5 °C), and reversal of any prior anticoagulation.

Outcomes

Primary Outcome — 6-month mortality

  • Adjusted HR 0.81 (95% CI 0.63–1.04); P=0.09 (NS).
  • Point estimate favors ≤2 h, but the trial was not powered to detect a mortality difference in this post-hoc surgical sub-cohort.

Secondary Outcomes (key, all adjusted):

  • Death or disability (mRS 4–6) at 6 months: adjusted OR 0.71 (95% CI 0.56–0.90); P=0.01.
  • mRS score shift (ordinal): adjusted common OR 0.73 (95% CI 0.60–0.89); P<0.01.
  • Serious adverse events (SAEs): adjusted OR 0.73 (95% CI 0.57–0.94); P=0.01.
  • Health-related quality of life (EQ-5D-3L): higher scores in mobility, pain/discomfort, and usual-activities domains in ≤2 h group; higher EQ-VAS and EQ-utility index.

Adverse Events / Safety:

  • “Fewer SAEs” in ≤2 h group without specification of SAE subtype breakdown; the directional safety signal is reassuring against concerns that rapid BP lowering in the perioperative window harms cerebral perfusion.

Figures

INTERACT3 surgical cohort study flowchart.
Figure 1. Study flowchart. INTERACT3 indicates the third Intensive Care Bundle With Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial; and SBP, systolic blood pressure.

Source: PMC PMC13196860 — str-57-1549-g001.jpg. Click image to expand.

Criticisms

  • Post-hoc subgroup of a stepped-wedge cluster-randomized trial. Exposure (time-to-target BP) was not randomized; the comparison is observational within a randomized parent trial. Residual confounding by indication (sicker patients may have taken longer to reach target) is the dominant threat to internal validity.
  • Time-zero definition is “hospital arrival,” not symptom onset. Delays in pre-hospital transport (LMIC settings) compress the operational window and may misclassify patients whose biological window had already closed.
  • Single BP cutoff (SBP <140 mm Hg) and binary 2-hour split. A continuous time-to-target or area-under-curve analysis over SBP would be more informative; dichotomization loses dose-response information.
  • No prespecified plan to adjust for hematoma volume, intraoperative blood loss, anesthetic agent, or BP-lowering drug class. Adjustment set is narrow (age, sex, treatment type, GCS).
  • Limited external validity to high-income settings — most sites were in LMIC stroke systems with longer pre-hospital delays, different anesthetic practice, and different surgical evacuation thresholds.
  • Mortality, the parent trial’s primary endpoint, is non-significant. All other benefits are secondary and adjusted; the headline “ultra-early BP” claim rests on ordinal-shift and SAE outcomes, which are more vulnerable to unmeasured confounding.
  • No core-lab adjudication of SBP trajectory or hematoma expansion at 24 h in this secondary analysis (the parent trial did adjudicate).

Funding

Award MR/T005009/1 jointly funded by the UK Department of Health and Social Care, the Foreign, Commonwealth & Development Office, the Medical Research Council, and the Wellcome Trust; the West China Hospital Outstanding Discipline Development 1-3-5 program (ZY2016102); the National Health and Medical Research Council of Australia (APP1149987); Sichuan Credit Pharmaceutical; and Takeda (China). The funders played roles in the parent INTERACT3 trial design; the present post-hoc analysis was investigator-initiated.

The paper

  • Authors. Liu L, Chen C, Liu J, Liu M, Ma L, Billot L, Li Q, Zhao Y, Malavera A, Muñoz-Venturelli P, de Silva DA, Nguyen HT, Wahab KW, Pandian JD, Wasay M, Pontes-Neto OM, Jiang Y, Abanto C, Arauz A, Song L, et al.
  • Title. Impact of Ultra-Early Perioperative Antihypertensive Therapy in Acute Intracerebral Hemorrhage.
  • Journal. Stroke.
  • Year. 2026.
  • DOI. 10.1161/STROKEAHA.125.053989
  • PMCID. PMC13196860
Deep Dive — click to expand

What this is

This is a post-hoc secondary analysis of INTERACT3, asking whether how fast you get an ICH patient to SBP <140 mm Hg matters specifically in the subgroup who go to the operating room for hematoma evacuation. The parent trial proved that the care bundle (BP, glucose, temperature, anticoagulation reversal) improves functional outcome; this paper slices the surgical sub-cohort by a 2-hour time-to-target cutoff. Headline result: faster BP control is associated with less death/disability, better mRS shift, fewer SAEs, and no mortality penalty — but the analysis is observational within a randomized trial, so “association” is the operative word.

1. Shadow Audit

The framing emphasizes a time-dependent treatment effect in surgical ICH, but the paper quietly buries three things:

  1. Mortality is non-significant (HR 0.81, P=0.09). The “ultra-early” headline is anchored on ordinal mRS shift (common OR 0.73) and SAE reduction (OR 0.73), which are softer endpoints. In an unblinded stepped-wedge bundle intervention, ordinal mRS shift is more vulnerable to assessment drift than mortality.
  2. The exposure is uncoupled from randomization. INTERACT3 randomized hospitals to the bundle, not to a speed of BP titration. Patients who reached target faster may be patients with smaller bleeds, less intraoperative bleeding, easier hemodynamics, or lower admission GCS — all of which independently predict outcome.
  3. The 2-hour cutoff is data-driven and dichotomized. A continuous time-to-target analysis (or restricted cubic spline) is absent from the abstract. Without it, the reader cannot tell whether the dose-response curve is monotonic, J-shaped, or flat until 6 h.

2. Inversion Engine

For the inverse conclusion (delayed BP control is fine, or even better, in surgical ICH) to hold, the authors would need:

  • The mortality benefit at ≤2 h to disappear once admission hematoma volume, intraoperative blood loss, and BP-lowering drug class are added to the model.
  • A competing-risk or U-shaped analysis showing that ultra-fast lowering transiently drops cerebral perfusion pressure below the lower limit of autoregulation in evacuated brains with disrupted vasculature.
  • A pre-specified, prospectively enrolled surgical subgroup from a different trial (e.g., a future ATACH-3 surgical arm) showing null or harm.

Quantitatively: the 29% relative reduction in death/disability (OR 0.71) needs to attenuate by ~50% to reach non-significance (OR ≈0.85). That is plausible if admission hematoma volume — a major unmeasured confounder — is added. The mortality HR would need to drift from 0.81 to ~0.95 or higher to neutralize.

3. Second-Order Catalyst

If real, the first adopters are:

  • Comprehensive stroke centers in LMIC settings (the trial’s actual geography) — they already titrate BP aggressively in non-surgical ICH; extending the protocol to the OR and PACU is a workflow change, not a new drug.
  • High-volume ICH evacuation centers in China and India where surgery is more liberally performed than in Western practice; the relative reduction applies most where the surgery happens most.
  • Anesthesia teams adopting a “stroke-BP-target handoff” so the post-induction SBP target (140 mm Hg) is met within 2 h of ED arrival, not 2 h post-op.

Speed of adoption: 6–18 months for protocol revision; 3–5 years for outcomes data to confirm or refute.

4. Asymmetric Leverage

  • Small effect, large denominator. OR 0.71 for death/disability across 1506 surgical patients means ≈430 cases annually could be prevented from severe disability globally if extrapolated — a moderate absolute number relative to total ICH burden (~3 million cases/year), but disproportionately impactful where surgical evacuation is the standard of care for moderate-volume supratentorial bleeds.
  • The asymmetric payoff is in functional outcome (mRS shift, HRQoL), not mortality. That means the patient-level benefit is felt in independence, return to work, and caregiver burden — not in headline survival rates. Payers and ICER-style HTA agencies often undervalue mRS-shift gains; bedside clinicians and families do not.
  • The cost asymmetry is favorable: nicardipine and labetalol are cheap, generic, and already on formulary. The marginal cost of “faster titration” is nursing attention, not new infrastructure.

5. Paradigm Destroyer

This paper kills the reflex that perioperative BP control can wait until after the craniotomy. The reflex was: “patient is asleep, the surgeon controls the field, we’ll titrate later.” Replace with: “SBP <140 mm Hg within 2 hours of ED arrival is a hard institutional metric, including the OR window; anesthesia sign-out documents time-to-target, not just intraoperative BP.”

Two-sentence change to a stroke protocol:

  1. Order an arterial line and an IV nicardipine/labetalol drip before craniotomy for ICH evacuation, with a target of SBP <140 mm Hg reached by 2 h after ED arrival.
  2. Track time-to-target as a quality metric alongside door-to-needle and door-to-groin-puncture times.

MVP — Minimum Viable Proof

“For a patient with acute ICH undergoing hematoma evacuation, achieving SBP <140 mm Hg within 2 hours of ED arrival is associated with a 29% lower odds of death or disability at 6 months (adjusted OR 0.71, 95% CI 0.56–0.90) without increasing serious adverse events.” If true at the bedside, it justifies prioritizing BP titration in parallel with the surgical booking.

Best Combination

Combine this finding with:

  • INTERACT3 main trial (NEJM 2023): the bundle lowers mRS-shift at 6 months in all ICH patients; this paper shows the BP component has a time-sensitive gradient specifically in the surgical sub-cohort.
  • ATACH-2 (NEJM 2016) and INTERACT pilot (2012): earlier intensive BP reduction is safe but did not improve mortality in non-surgical ICH. The present surgical sub-analysis is consistent with a paradigm where surgical + early-BP is the high-leverage combination.
  • PREDICT/Sunnybrook spot-sign data: hematoma expansion is highest in the first 2–3 h; ultra-early BP control plausibly interrupts that expansion curve.
  • 2022 AHA/ASA ICH guideline endorses early BP lowering to <140 mm Hg; this paper supplies the time-target rationale that guideline currently lacks.

Overvalue Warning

  1. Do not over-read this as proof that ultra-early BP control reduces mortality. Mortality HR is 0.81, P=0.09 — not significant. The benefit is functional and HRQoL, not survival.
  2. Do not extrapolate from the INTERACT3 surgical subgroup to high-income settings without pre-hospital delays. The exposure window starts at “hospital arrival,” which in this dataset is often hours after symptom onset in LMIC contexts; in a Western EMS-fast system the analogous cutoff might be 30–60 minutes, and the dose-response curve at that end is unknown.

Educational use notice. The Clotbust Review is dedicated to medical education. Figures from published articles are embedded for teaching purposes, with attribution to the original publication. We do not claim copyright over any embedded figure; all rights remain with the original publisher and authors under the article's published license. If you are a rights holder and would like a figure removed, please contact the editorial team. See our editorial notice for the full policy.