Publication

• Title: Rapid blood-pressure lowering in patients with acute intracerebral hemorrhage
• Acronym: INTERACT2 (Second Intensive Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial)
• Year: 2013
• Journal published in: New England Journal of Medicine
• Citation: Anderson CS, Heeley E, Huang Y, Wang J, Stapf C, Delcourt C, et al; INTERACT2 Investigators. Rapid blood-pressure lowering in patients with acute intracerebral hemorrhage. N Engl J Med. 2013;368(25):2355-65.

Context & Rationale

• Background

  • Spontaneous intracerebral haemorrhage (ICH) has high early mortality and long-term disability, with few proven acute medical therapies.
  • Haematoma expansion occurs early after ICH and is strongly associated with worse outcomes, making it an attractive modifiable target.
  • Markedly elevated blood pressure (BP) is common after ICH and is associated with haematoma expansion and poor outcome; however, lowering BP raised concerns about cerebral hypoperfusion (especially in patients with chronic hypertension and/or raised intracranial pressure).
  • Prior practice guidelines were conservative (often targeting systolic BP <180 mm Hg), reflecting limited randomised evidence and equipoise.
  • The pilot INTERACT1 study suggested feasibility and safety of intensive BP lowering to a systolic target of 140 mm Hg and a signal towards reduced haematoma growth, motivating a definitive outcomes trial.
• Research Question/Hypothesis

  • In adults with acute spontaneous ICH within 6 hours and systolic BP 150–220 mm Hg, does early intensive systolic BP lowering to <140 mm Hg within 1 hour (maintained for 7 days) reduce death or major disability at 90 days compared with guideline-recommended BP management targeting systolic BP <180 mm Hg?
• Why This Matters

  • BP lowering is inexpensive, widely deployable, and immediately actionable across emergency, stroke-unit, and ICU settings.
  • Even modest functional benefits would have large population-level impact given the global burden of ICH.
  • Safety data are critical because overly aggressive BP reduction could plausibly worsen brain and end-organ perfusion.

Design & Methods

• Research Question: Whether a management strategy of rapid intensive systolic BP lowering to <140 mm Hg improves 90-day clinical outcome after acute spontaneous ICH compared with guideline-recommended management targeting systolic BP <180 mm Hg.
• Study Type: International, multicentre, prospective, randomised, open-treatment with blinded end-point assessment (PROBE); conducted in acute stroke/critical care-capable hospital settings (144 hospitals, 21 countries).
• Population:
  • Inclusion: Adults with spontaneous intracranial haemorrhage confirmed by CT/MRI; systolic BP 150–220 mm Hg; randomisation and commencement of BP-lowering management feasible within 6 hours of symptom onset.
  • Key exclusions: Structural cause for ICH; deep coma (GCS 3–5); massive haematoma with very poor prognosis; planned early surgical evacuation; definite indication for, or contraindication to, protocolised BP lowering.
• Intervention:
  • Early intensive BP lowering with target systolic BP <140 mm Hg within 1 hour of randomisation, maintained for 7 days.
  • Locally available intravenous antihypertensive agents using prespecified, site-specific dose-escalation protocols; transition to oral agents when feasible; delivered in a monitored environment.
• Comparison:
  • Guideline-recommended BP management with target systolic BP <180 mm Hg using locally available agents and usual clinical care processes in monitored settings.
• Blinding: Open-label treatment delivery; blinded end-point assessment at 90 days (masked outcome assessors).
• Statistics: A sample size of ~2800 participants was planned to detect a 14% relative reduction in the primary event rate (from 50% to 43%) with 90% power at the 5% significance level (allowing ~10% non-adherence and ~3% loss to follow-up); primary analyses were intention-to-treat; ordinal logistic regression was prespecified for the distributional (shift) analysis of modified Rankin Scale (mRS) outcomes; two interim analyses used a Haybittle–Peto stopping approach (nominal P<0.048 for final analyses).
• Follow-Up Period: 90 days for primary and key secondary clinical outcomes; BP separation tracked from randomisation to day 7; repeat 24-hour CT imaging in a prespecified subgroup.

Key Results

This trial was not stopped early. Two interim analyses were planned and conducted with conservative stopping boundaries; recruitment continued to the full sample.

• BP separation was achieved but incomplete: mean systolic BP at 1 hour was 150 mm Hg (intensive) vs 164 mm Hg (guideline), a 14 mm Hg difference; 462/1382 (33.4%) achieved <140 mm Hg at 1 hour.
• The primary dichotomous end point narrowly missed conventional statistical significance, while the prespecified ordinal analysis showed a small but favourable shift in functional outcome.
• Major safety signals were not seen in INTERACT2 (severe hypotension <1% in both groups), though later trials and pooled analyses refined “how low” is safe and beneficial (see Controversies & Subsequent Evidence).

Internal Validity

• Randomisation and allocation concealment: Central secure web-based randomisation with minimisation to balance country, hospital, and time from onset (≤4 hours vs >4 hours), supporting allocation concealment and baseline balance.
• Dropout / exclusions (post-randomisation): Primary outcome determined for 1382/1403 (98.5%) in the intensive group and 1412/1436 (98.3%) in the guideline group; missingness was mainly due to loss to follow-up, withdrawal of consent, or missing mRS data at 90 days.
• Performance/detection bias: Open-label delivery creates risk of differential co-interventions; end-point assessment was masked, reducing detection bias for mRS.
• Protocol adherence and contamination: Intensive targets were not consistently achieved early; 462/1382 (33.4%) achieved systolic BP <140 mm Hg at 1 hour; 26.6% vs 8.1% received ≥2 intravenous agents (intensive vs guideline), indicating attempted intensification but limited early target attainment.
• Baseline comparability and illness severity: Groups were reported as well balanced at baseline; exclusions (deep coma and massive haematoma) removed the sickest tail, focusing inference on patients for whom medical management plausibly alters outcome.
• Heterogeneity: Broad international enrolment and locally selected antihypertensive regimens increase clinical heterogeneity; prespecified subgroup analyses showed no convincing effect modification (all interaction P values non-significant), though subgroup analyses were underpowered for modest interactions.
• Timing: Median onset-to-initiation of IV treatment was 4.0 hours (IQR 2.9–5.1) vs 4.5 hours (IQR 3.0–7.0) (intensive vs guideline); randomisation-to-treatment initiation was 6 minutes (IQR 0–39) vs 19 minutes (IQR 0–167), consistent with earlier treatment in the intensive arm.
• Dose: The intended “dose” (rapidly reaching <140 within 1 hour and sustaining for 7 days) was only partially delivered; limited early target attainment plausibly diluted efficacy.
• Separation of the variable of interest: Mean systolic BP at 1 hour was 150 mm Hg vs 164 mm Hg (difference 14 mm Hg; P<0.001); at 6 hours, mean systolic BP was approximately 139 mm Hg vs 153 mm Hg (difference 14 mm Hg), with significant between-group separation sustained to day 7.
• Key delivery aspects: Separation achieved early and sustained, but the absolute BP contrast and “time-at-target” were modest relative to protocol aspiration, consistent with pragmatic execution constraints across 144 sites.
• Adjunctive therapy use: Most co-interventions during the first 7 days were similar between groups; however, withdrawal of active treatment was more frequent in the intensive group (75/1399 [5.4%] vs 46/1430 [3.3%]).
• Outcome assessment: Primary outcome used standardised mRS at 90 days with masked assessment; mortality is objective; repeat CT imaging outcomes were confined to a prespecified subgroup (491 vs 473), introducing potential selection effects for radiological analyses.
• Statistical rigour: Intention-to-treat framework with prespecified interim monitoring and adjusted significance threshold (nominal P<0.048); the primary dichotomous outcome narrowly missed significance (P=0.06) while the prespecified ordinal analysis was nominally significant (P=0.04).

Conclusion on Internal Validity: Overall, internal validity appears moderate: allocation procedures and outcome masking were robust, but open-label care and incomplete early target attainment (limited BP “dose” separation) plausibly diluted treatment effects and introduce residual performance bias.

External Validity

• Population representativeness: Large, pragmatic enrolment (2839 participants; 144 hospitals; 21 countries) supports generalisability to many acute stroke/critical care systems, though recruitment was regionally weighted (substantial enrolment from China) and excluded very severe presentations (deep coma, massive haematoma) and planned early surgery.
• Applicability: Results are most applicable to adults with spontaneous ICH presenting within 6 hours with systolic BP 150–220 mm Hg in settings able to deliver monitored IV BP control; applicability is less certain for patients with extreme hypertension (>220 mm Hg), those requiring emergent neurosurgery, or patients with marked raised intracranial pressure where individualised haemodynamic targets may be needed.
• Intervention transportability: Use of locally available antihypertensive agents increases real-world relevance across health systems but also means the effect represents a strategy rather than a single agent regimen.

Conclusion on External Validity: External validity is moderate to strong for typical hyperacute, non-surgical spontaneous ICH managed in monitored units, but limited for the extremes of severity and for settings lacking the capacity for rapid, titrated IV BP control.

Strengths & Limitations

• Strengths:
  • Large, international, pragmatic RCT with high follow-up completeness for the primary outcome.
  • Clinically meaningful patient-centred primary endpoint at 90 days with masked outcome assessment.
  • Prespecified ordinal (shift) analysis capturing distributional functional effects beyond dichotomisation.
  • Demonstrated feasibility and safety of targeting systolic BP <140 mm Hg across diverse settings.
• Limitations:
  • Open-label care delivery with potential for performance bias and differential decisions (including withdrawal of active care).
  • Incomplete achievement of the intended BP target early after randomisation (only 33.4% reached <140 mm Hg at 1 hour), potentially diluting efficacy.
  • Radiological mechanistic outcomes (haematoma growth) assessed in a subgroup, limiting mechanistic inference and increasing vulnerability to selection effects.
  • Effect estimates were modest and near the margin of statistical significance for key outcomes, heightening sensitivity to analytic choices (dichotomous vs ordinal) and to multiplicity considerations.

Interpretation & Why It Matters

• Clinical signal

  In hyperacute spontaneous ICH with systolic BP 150–220 mm Hg, a strategy targeting systolic BP <140 mm Hg was safe and associated with a favourable shift in 90-day functional outcomes (ordinal OR 0.87; 95% CI 0.77 to 1.00; P=0.04), despite a non-significant primary dichotomous end point (OR 0.87; 95% CI 0.75 to 1.01; P=0.06).
• Implementation nuance

  Benefits, if real, are likely contingent on speed and smoothness of BP control: mean systolic BP was 150 vs 164 mm Hg at 1 hour, and only 33.4% achieved <140 mm Hg at 1 hour, indicating that real-world performance may lag protocol aspiration.
• Mechanism and expectations

  Absence of a clear between-group reduction in 24-hour haematoma growth in the repeat-CT subgroup (adjusted mean difference 1.4 mL; 95% CI −0.6 to 3.4; P=0.18) suggests that any functional benefit may be mediated by pathways beyond measurable haematoma expansion in this subset, or that treatment intensity/timing was insufficient to reliably alter bleeding.

Controversies & Other Evidence

• Primary dichotomous outcome vs ordinal shift: The trial’s headline conclusion depends on analytic framing: dichotomised death/major disability narrowly missed significance (P=0.06), whereas the prespecified ordinal analysis reached nominal significance (P=0.04), raising interpretive sensitivity to endpoint selection and multiplicity.
• Safety and “how low to go”: Contemporary editorial commentary emphasised that cerebral perfusion risk (particularly in patients with chronic hypertension and potential intracranial hypertension) is biologically plausible and that cautious, monitored lowering is essential when extrapolating strategy-level trial findings to heterogeneous real-world patients.¹
• Agent heterogeneity and physiological effects: Correspondence around INTERACT2 highlighted that the intervention permitted multiple BP-lowering drug classes with differing haemodynamic and cerebrovascular effects, complicating mechanistic inference about whether benefits (or harms) are strategy-, agent-, or implementation-dependent.²
• Concerns about excessive lowering (J-shaped association): Subsequent correspondence discussing intensive BP lowering after ICH referenced data suggesting a potential J-shaped relationship between achieved systolic BP and outcome, supporting guideline caution against very low targets in unselected patients.³
• Subsequent large RCT with lower targets (ATACH-2): ATACH-2 compared more intensive targets (110–139 mm Hg) with standard targets (140–179 mm Hg) using nicardipine and was stopped early for futility; there was no improvement in death/disability (38.7% vs 37.7%), while renal adverse events within 7 days were higher with intensive lowering (9.0% vs 4.0%; P=0.002), sharpening the message that “lower is not necessarily better”.⁴
• Individual patient data meta-analysis (BASC): A one-stage IPD meta-analysis found no overall functional benefit of BP-lowering interventions after ICH (adjusted OR for unfavourable mRS shift 0.97; 95% CI 0.88 to 1.06; p=0.50) despite clear reductions in haematoma growth (absolute >6 mL: OR 0.75; 95% CI 0.60 to 0.92; p=0.0077; relative ≥33%: OR 0.82; 95% CI 0.68 to 0.99; p=0.034), with heterogeneity by strategy and agent.⁵
• BP variability and quality of control: Post-hoc INTERACT2 analyses associated greater systolic BP variability with poorer outcomes, supporting an implementation focus on smooth, sustained BP control and avoidance of peaks rather than single time-point targets alone.⁶
• Systems-level implementation (INTERACT3): A stepped-wedge cluster randomised trial of a goal-directed acute care bundle (including early SBP target <140 mm Hg plus correction of fever, hyperglycaemia, and warfarin-associated coagulopathy when present) improved 6-month functional outcome (common OR 0.86; 95% CI 0.76–0.97; p=0.015) and reduced serious adverse events (16.0% vs 20.1%; p=0.0098), suggesting that BP control may be most effective when embedded within broader physiological optimisation.⁷
• Guidelines incorporating INTERACT2 and later evidence: Major contemporary guidelines generally support early lowering of systolic BP towards ~140 mm Hg in eligible patients with acute ICH while avoiding overly aggressive reduction to very low targets, reflecting the balance of INTERACT2’s safety/functional signal and ATACH-2’s renal safety concerns.⁸⁹
• Further Reading
  • Trials
    • Qureshi AI, Palesch YY, Barsan WG, Hanley DF, Hsu CY, Martin RL, et al; ATACH-2 Trial Investigators and the Neurological Emergency Treatment Trials Network. Intensive blood-pressure lowering in patients with acute cerebral hemorrhage. N Engl J Med. 2016;375(11):1033-43. DOI
    • Ma L, Hu X, Song L, et al; INTERACT3 Investigators. The third Intensive Care Bundle with Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial (INTERACT3): an international, stepped wedge cluster randomised controlled trial. Lancet. 2023;402(10395):27-40. DOI
  • Meta-analyses and pooled analyses
    • Moullaali TJ, Wang X, Sandset EC, Woodhouse LJ, Law ZK, Arima H, et al; Blood Pressure in Acute Stroke (BASC) Investigators. Early lowering of blood pressure after acute intracerebral haemorrhage: a systematic review and meta-analysis of individual patient data. J Neurol Neurosurg Psychiatry. 2022;93(1):6-13. DOI
  • Observational / post-hoc analyses
    • Manning L, Hirakawa Y, Arima H, Wang X, Chalmers J, Wang J, et al; INTERACT2 Investigators. Blood pressure variability and outcome after acute intracerebral haemorrhage: a post-hoc analysis of INTERACT2, a randomised controlled trial. Lancet Neurol. 2014;13(4):364-73. DOI
  • Guidelines
    • Greenberg SM, Ziai WC, Cordonnier C, et al. 2022 Guideline for the Management of Patients With Spontaneous Intracerebral Hemorrhage: A Guideline From the American Heart Association/American Stroke Association. Stroke. 2022;53(7):e282-e361. DOI
    • Sandset EC, Anderson CS, Bath PM, Christensen H, Fischer U, Gąsecki D, et al. European Stroke Organisation (ESO) guidelines on blood pressure management in acute ischaemic stroke and intracerebral haemorrhage. Eur Stroke J. 2021;6(2):XLVIII-LXXXIX. DOI

Summary

• INTERACT2 randomised 2839 patients with hyperacute spontaneous ICH (≤6 hours) and systolic BP 150–220 mm Hg to intensive (<140) vs guideline (<180) systolic BP targets.
• The primary outcome (death or major disability, mRS 3–6 at 90 days) was lower with intensive treatment (52.0% vs 55.6%) but did not reach statistical significance (OR 0.87; 95% CI 0.75 to 1.01; P=0.06).
• The prespecified ordinal analysis showed a favourable shift in functional outcome (OR 0.87; 95% CI 0.77 to 1.00; P=0.04), with improved EQ-5D utility (0.60 ± 0.39 vs 0.55 ± 0.40; P=0.002).
• Safety outcomes were broadly similar (severe hypotension 0.5% vs 0.6%; nonfatal SAEs 23.3% vs 23.6%), supporting feasibility and short-term safety of targeting ~140 in monitored settings.
• Subsequent trials and pooled analyses indicate that very aggressive targets (e.g., 110–139) may increase renal adverse events without improving outcomes, reinforcing guideline emphasis on early, controlled lowering rather than “as low as possible”.

Overall Takeaway

INTERACT2 established that early BP lowering towards a systolic target of 140 mm Hg in eligible patients with hyperacute spontaneous ICH is practicable and safe across diverse health systems, and it suggested a modest but clinically relevant improvement in functional outcome when analysed across the full disability distribution. While the primary dichotomous endpoint narrowly missed significance, the trial helped shift practice and guidelines towards earlier, titrated BP control—tempered by later evidence that overly aggressive lowering can introduce renal harm without additional neurological benefit.

Bibliography

• 1.Frontera JA. Blood Pressure in Intracerebral Hemorrhage—How Low Should We Go? N Engl J Med. 2013;368(25):2426-7.
• 2.Keller DL, Vidale S, Mulder M, Geocadin RG, Naval NS, Frontera JA, Anderson CS, Chalmers J, Stapf C. Blood-pressure lowering in acute intracerebral hemorrhage. N Engl J Med. 2013;369(13):1273-5.
• 3.Frontera JA, Rabinstein AA. Intensive Blood-Pressure Lowering in Cerebral Hemorrhage. N Engl J Med. 2016;375(23):e48.
• 4.Qureshi AI, Palesch YY, Barsan WG, Hanley DF, Hsu CY, Martin RL, et al; ATACH-2 Trial Investigators and the Neurological Emergency Treatment Trials Network. Intensive blood-pressure lowering in patients with acute cerebral hemorrhage. N Engl J Med. 2016;375(11):1033-43.
• 5.Moullaali TJ, Wang X, Sandset EC, Woodhouse LJ, Law ZK, Arima H, et al; Blood Pressure in Acute Stroke (BASC) Investigators. Early lowering of blood pressure after acute intracerebral haemorrhage: a systematic review and meta-analysis of individual patient data. J Neurol Neurosurg Psychiatry. 2022;93(1):6-13.
• 6.Manning L, Hirakawa Y, Arima H, Wang X, Chalmers J, Wang J, et al; INTERACT2 Investigators. Blood pressure variability and outcome after acute intracerebral haemorrhage: a post-hoc analysis of INTERACT2, a randomised controlled trial. Lancet Neurol. 2014;13(4):364-73.
• 7.Ma L, Hu X, Song L, et al; INTERACT3 Investigators. The third Intensive Care Bundle with Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial (INTERACT3): an international, stepped wedge cluster randomised controlled trial. Lancet. 2023;402(10395):27-40.
• 8.Greenberg SM, Ziai WC, Cordonnier C, et al. 2022 Guideline for the Management of Patients With Spontaneous Intracerebral Hemorrhage: A Guideline From the American Heart Association/American Stroke Association. Stroke. 2022;53(7):e282-e361.
• 9.Sandset EC, Anderson CS, Bath PM, Christensen H, Fischer U, Gąsecki D, et al. European Stroke Organisation (ESO) guidelines on blood pressure management in acute ischaemic stroke and intracerebral haemorrhage. Eur Stroke J. 2021;6(2):XLVIII-LXXXIX.