Skip to main content

Foundational Trials

CATIS

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Effects of Immediate Blood Pressure Reduction on Death and Major Disability in Patients With Acute Ischemic Stroke: The CATIS Randomized Clinical Trial
  • Acronym: CATIS
  • Year: 2014
  • Journal published in: JAMA
  • Citation: He J, Zhang Y, Xu T, et al; CATIS Investigators. Effects of Immediate Blood Pressure Reduction on Death and Major Disability in Patients With Acute Ischemic Stroke: The CATIS Randomized Clinical Trial. JAMA. 2014;311(5):479-489.

Context & Rationale

  • Background
    • Elevated blood pressure (BP) is common in the first hours–days after acute ischaemic stroke and is associated (observationally) with both favourable and unfavourable outcomes depending on timing, stroke subtype, and reperfusion status.
    • Physiologically, acute hypertension may be compensatory (maintaining collateral/penumbral perfusion), yet it may also promote cerebral oedema, haemorrhagic transformation, and cardiac complications.
    • Prior trials of early BP modulation in acute stroke were limited by small sample sizes, heterogeneity of interventions, mixed stroke types, and inconsistent signals for benefit or harm.
    • A common bedside dilemma remained unresolved: whether to actively lower BP early after ischaemic stroke versus temporarily withholding chronic antihypertensives until the patient is clinically stable.
  • Research Question/Hypothesis
    • In adults with CT/MRI-confirmed acute ischaemic stroke within 48 hours and elevated BP, does immediate antihypertensive treatment (targeting a modest early SBP reduction and later attainment of <140/90 mm Hg) reduce death or major disability compared with temporary discontinuation of antihypertensive drugs during hospitalisation?
  • Why This Matters
    • Acute BP management decisions occur daily in stroke units, emergency departments, and critical care environments, and influence competing risks (cerebral perfusion vs haemorrhage/oedema/cardiac strain).
    • A large pragmatic RCT with robust BP separation and blinded endpoint assessment was needed to move beyond physiology-based practice variation.
    • Findings were directly positioned to inform guideline thresholds and the timing of restarting (or holding) chronic antihypertensive therapy after ischaemic stroke.

Design & Methods

  • Research Question: Among adults with acute ischaemic stroke (≤48 hours) and elevated BP, does immediate BP reduction improve clinical outcomes compared with temporarily withholding antihypertensive therapy during hospitalisation?
  • Study Type: Multicentre, randomised, single-blind, blinded end-point trial conducted in 26 hospitals in China (hospital stroke care setting).
  • Population:
    • Key inclusion: Age ≥22 years; acute ischaemic stroke confirmed by CT/MRI; randomisation within 48 hours of symptom onset; systolic BP 140 to <220 mm Hg.
    • Key exclusions (selected): Systolic BP ≥220 mm Hg or diastolic BP ≥120 mm Hg; receipt of intravenous thrombolysis; major comorbid cardiovascular instability (e.g., acute myocardial infarction/unstable angina, heart failure); deep coma; other conditions judged to mandate urgent BP treatment.
  • Intervention:
    • Initial management: Discontinuation of pre-stroke antihypertensive medications, then protocolised antihypertensive treatment started immediately after randomisation.
    • Targets: Reduce systolic BP by 10%–25% within the first 24 hours; achieve systolic/diastolic BP <140/<90 mm Hg within 7 days and maintain during hospitalisation.
    • Delivery: Stepwise, centre-protocol approach using commonly available antihypertensive agents (with dose titration and safety limits to avoid excessive BP reduction).
  • Comparison:
    • Initial management: Discontinuation of pre-stroke antihypertensive medications, then no antihypertensive treatment during hospitalisation unless clinically required outside the trial framework.
    • Co-interventions: Otherwise standard acute ischaemic stroke care per local practice/guidance; antihypertensive therapy after discharge at clinician discretion.
  • Blinding: Treating clinicians were unblinded; patients and outcome assessors were blinded; outcomes were adjudicated by a blinded end-point committee (single-blind with blinded end-points).
  • Statistics: A total of 4000 participants were planned to detect a 5% absolute reduction in death or major disability (from 35% to 30%) with 90% power at the 5% two-sided significance level (allowing ~8% non-adherence or loss to follow-up); primary analyses were intention-to-treat.
  • Follow-Up Period: 14 days or hospital discharge (whichever came first) and 3 months post-randomisation.

Key Results

This trial was not stopped early.

Outcome Antihypertensive treatment Control Effect p value / 95% CI Notes
Systolic BP at 24 h, mean (SD), mm Hg 144.7 (15.0) 152.9 (15.9) Mean difference −8.1 95% CI −9.1 to −7.2; P<.001 Demonstrates early physiological separation
Systolic BP at day 7, mean (SD), mm Hg 137.3 (11.8) 146.5 (13.6) Mean difference −9.3 95% CI −10.1 to −8.4; P<.001 Targeting <140/<90 by day 7
Systolic BP at day 14, mean (SD), mm Hg 135.2 (10.4) 143.7 (14.0) Mean difference −8.6 95% CI −9.7 to −7.4; P<.001 Sustained in-hospital BP separation
Primary: death or major disability (mRS 3–6) at 14 days/discharge, n (%) 683/2038 (33.6) 681/2033 (33.6) OR 1.00 95% CI 0.88 to 1.14; P=.98 Primary end point (neutral)
mRS shift (odds of 1-unit higher mRS) at 14 days/discharge — — OR 0.98 95% CI 0.88 to 1.09; P=.70 Ordinal analysis (neutral)
Death at 14 days/discharge, n (%) 25/2038 (1.2) 25/2033 (1.2) OR 1.00 95% CI 0.57 to 1.74; P=.99 Low early mortality
Systolic BP at 3 months, mean (SD), mm Hg 139.3 (11.7) 142.2 (2.5) Mean difference −2.9 95% CI −3.7 to −2.2 Between-group separation narrowed by 3 months
Antihypertensive medication use at 3 months, n (%) 1680/1988 (84.5) 1499/1987 (75.4) OR 1.78 95% CI 1.52 to 2.09 More patients on BP therapy at follow-up in intervention arm
Death or major disability (mRS 3–6) at 3 months, n (%) 500/1988 (25.2) 502/1987 (25.3) OR 0.99 95% CI 0.86 to 1.15; P=.93 Sustained neutrality at 90 days
Death at 3 months, n (%) 68/1988 (3.4) 54/1987 (2.7) OR 1.27 95% CI 0.88 to 1.82; P=.20 No statistically significant difference
Recurrent stroke at 3 months, n (%) 28/1988 (1.4) 43/1987 (2.2) OR 0.65 95% CI 0.40 to 1.04; P=.07 Numerical reduction; not statistically significant
Prespecified subgroup: time from onset to randomisation ≥24 h — death/major disability at 3 months, n (%) 108/597 (18.1) 139/600 (23.2) OR 0.73 95% CI 0.55 to 0.97; P=.03 Interaction (P for homogeneity .03) suggests time-dependence; interpret cautiously
  • Protocolised antihypertensive treatment achieved clinically meaningful early BP separation (e.g., SBP 144.7±15.0 vs 152.9±15.9 mm Hg at 24 hours; mean difference −8.1 mm Hg; 95% CI −9.1 to −7.2; P<.001) without improvement in the primary clinical outcome (33.6% vs 33.6%; OR 1.00; 95% CI 0.88 to 1.14; P=.98).
  • At 3 months, SBP remained lower in the intervention arm (139.3±11.7 vs 142.2±2.5 mm Hg; mean difference −2.9; 95% CI −3.7 to −2.2), yet death/major disability remained unchanged (25.2% vs 25.3%; OR 0.99; 95% CI 0.86 to 1.15; P=.93).
  • A prespecified interaction by time-to-randomisation suggested possible benefit when treatment began ≥24 hours after onset (OR 0.73; 95% CI 0.55 to 0.97; P=.03), not observed earlier.

Internal Validity

  • Randomisation and Allocation:
    • Randomisation was central and stratified by hospital and pre-stroke antihypertensive use, supporting allocation concealment and minimising selection bias.
    • Baseline balance was strong for key prognostic variables (e.g., entry SBP 166.7±17.3 vs 165.6±16.5 mm Hg; NIHSS median 4.0 [IQR 2.0–7.0] vs 4.0 [IQR 3.0–8.0]).
  • Drop out or exclusions (post-randomisation):
    • Randomised: 2038 (intervention) vs 2033 (control).
    • 3-month follow-up cohort in Table 3: 1988 vs 1987 (≈50 and 46 participants without 3-month outcome data, respectively).
    • Protocol specified withdrawal criteria for severe clinical deterioration or persistent extreme hypertension; numbers withdrawn for protocol-defined safety reasons were not reported.
  • Performance/Detection Bias:
    • Treating clinicians were not blinded, which could influence co-interventions and discharge practices.
    • Outcome assessment was blinded and endpoints were adjudicated, reducing detection bias for mRS-based outcomes.
  • Protocol Adherence:
    • The intervention achieved sustained BP separation across the in-hospital period, suggesting high protocol fidelity.
    • Control contamination occurred: 2.3% of control participants did not discontinue home antihypertensives.
  • Baseline Characteristics:
    • Stroke severity was generally mild-to-moderate (NIHSS median 4), which can reduce sensitivity to detect treatment effects that primarily influence large infarcts or malignant oedema trajectories.
    • Mean age was 62.0±10.2 years vs 62.2±9.8 years; 64.4% vs 65.3% were men.
  • Heterogeneity:
    • Prespecified subgroup analyses did not show effect modification at 14 days/discharge.
    • A time-to-randomisation interaction at 3 months (P for homogeneity .03) raises the possibility of time-dependent effects, but remains hypothesis-generating given multiplicity.
  • Timing:
    • Randomisation occurred within 48 hours, with mean onset-to-randomisation times 15.3±12.9 hours vs 14.9±13.0 hours.
    • This timing tests “early” rather than “hyperacute” BP lowering, which is clinically important for interpreting generalisability to prehospital or reperfusion settings.
  • Dose:
    • The intended “dose” was a moderate SBP reduction (10%–25% within 24 hours) with safety limits to avoid excessive hypotension; the achieved SBP reduction at 24h was −21.8±17.7 vs −12.7±16.3 mm Hg (difference −9.1; 95% CI −10.2 to −8.1; P<.001).
  • Separation of the Variable of Interest:
    • SBP at 24 h: 144.7±15.0 vs 152.9±15.9 mm Hg (difference −8.1; 95% CI −9.1 to −7.2; P<.001).
    • SBP at day 7: 137.3±11.8 vs 146.5±13.6 mm Hg (difference −9.3; 95% CI −10.1 to −8.4; P<.001).
    • SBP at day 14: 135.2±10.4 vs 143.7±14.0 mm Hg (difference −8.6; 95% CI −9.7 to −7.4; P<.001).
  • Key Delivery Aspects:
    • The control strategy was explicit “temporary discontinuation” rather than “usual care with permissive treatment,” which strengthens internal contrast but affects interpretability across health systems.
    • High rates of background antithrombotic/heparin use were broadly balanced and unlikely to fully account for neutral outcomes.
  • Outcome Assessment:
    • Primary endpoint used a clinically meaningful dichotomy (mRS 3–6) at 14 days/discharge; 3-month mRS provided a standard longer-term functional assessment.
  • Statistical Rigor:
    • The primary result was consistent in adjusted sensitivity analyses (e.g., adjusted OR 1.09; 95% CI 0.92 to 1.28; P=.33 at 14 days/discharge), supporting robustness.
    • The sample size target was met (n=4071), with event rates near assumptions used for power.

Conclusion on Internal Validity: Overall, internal validity appears moderate-to-strong given central randomisation, balanced baseline characteristics, blinded endpoint assessment, sustained BP separation, and low attrition; however, unblinded treatment delivery and minor control contamination could have influenced co-interventions.

External Validity

  • Population Representativeness:
    • Represents hospital-treated, predominantly mild-to-moderate acute ischaemic stroke with elevated BP (SBP 140–<220) in a Chinese multicentre setting.
    • Excludes key contemporary stroke populations: patients receiving intravenous thrombolysis, those with extreme hypertension (≥220/≥120), and patients with major cardiovascular instability or deep coma.
  • Applicability:
    • Most applicable to non-reperfused acute ischaemic stroke patients where clinicians consider either permissive hypertension or early modest BP reduction.
    • Less directly applicable to hyperacute reperfusion pathways (thrombolysis/thrombectomy) where BP thresholds are protocolised and haemorrhagic transformation risk is a competing concern.
    • Background care (e.g., anticoagulant/heparin use rates and length of stay) may differ materially from many European health systems, potentially influencing absolute event rates.

Conclusion on External Validity: Generalisability is moderate: CATIS informs early BP lowering versus temporary withholding in non-thrombolysed acute ischaemic stroke with SBP 140–<220, but extrapolation to reperfusion-treated, hyperacute, or very severe stroke cohorts should be cautious.

Strengths & Limitations

  • Strengths:
    • Large pragmatic multicentre RCT (n=4071) with clinically important endpoints and high follow-up completeness.
    • Clear and sustained BP separation between groups (≈8–9 mm Hg in-hospital SBP differences).
    • Blinded outcome assessment and adjudication for functional endpoints.
    • Explicit comparison of an “active BP-lowering strategy” versus “temporary discontinuation” of antihypertensive drugs, addressing a common real-world decision.
  • Limitations:
    • Exclusion of thrombolysis limits relevance to modern reperfusion-heavy pathways.
    • Mean onset-to-randomisation time ~15 hours means the trial does not answer hyperacute (prehospital/first hours) BP-lowering questions.
    • Open-label treatment delivery may influence co-interventions and in-hospital management despite blinded endpoints.
    • Stroke severity was generally mild-to-moderate (NIHSS median 4), which may reduce ability to detect effects in severe stroke phenotypes where BP manipulation could plausibly matter more.

Interpretation & Why It Matters

  • Clinical inference (non-reperfused AIS)
    In hospital-treated acute ischaemic stroke with SBP 140–<220 mm Hg, immediate protocolised BP reduction (10%–25% in 24 h, then <140/<90 by 7 days) did not improve death or disability at either 14 days/discharge (33.6% vs 33.6%) or 3 months (25.2% vs 25.3%).
  • Practical management implication
    For similar patients, routine early antihypertensive escalation aimed at modest BP reduction cannot be justified solely to improve functional outcome; temporary withholding of chronic antihypertensives during the earliest inpatient phase appears clinically reasonable when BP is below severe thresholds and there are no competing indications.
  • Methodological implication
    The neutral result despite robust BP separation suggests that (in this population and time-window) BP level itself may be an epiphenomenon of stroke physiology rather than a modifiable determinant of early functional outcome; future trials must interrogate timing, stroke mechanism, imaging-defined penumbra/collaterals, and reperfusion status as key effect modifiers.

Controversies & Subsequent Evidence

  • Physiological concern persists that early BP reduction may compromise collateral perfusion in vulnerable penumbral tissue; CATIS primarily evaluates an “early but not hyperacute” window (mean ~15 hours), leaving the first hours after onset less directly informed.1
  • Systematic reviews of randomised BP-lowering strategies in early ischaemic stroke similarly report no clear improvement in functional outcome overall, aligning with CATIS’s neutrality and reinforcing the importance of clinical context (timing, reperfusion, and severity) when interpreting BP signals.2
  • Guidelines published after CATIS generally recommend against routine BP lowering in acute ischaemic stroke unless BP is severely elevated or there are compelling comorbid indications, and specify different BP thresholds for patients undergoing thrombolysis/thrombectomy—recommendations consistent with CATIS’s findings in non-reperfused stroke.34

Summary

  • CATIS randomised 4071 patients with acute ischaemic stroke (≤48 h) and elevated BP to immediate protocolised antihypertensive therapy versus temporary discontinuation of antihypertensives during hospitalisation.
  • The intervention achieved robust in-hospital BP separation (e.g., SBP 144.7±15.0 vs 152.9±15.9 mm Hg at 24 h; mean difference −8.1; 95% CI −9.1 to −7.2; P<.001).
  • There was no difference in the primary endpoint at 14 days/discharge (33.6% vs 33.6%; OR 1.00; 95% CI 0.88 to 1.14; P=.98) and no benefit at 3 months (25.2% vs 25.3%; OR 0.99; 95% CI 0.86 to 1.15; P=.93).
  • Early mortality was low and identical at 14 days/discharge (1.2% vs 1.2%); 3-month mortality did not differ significantly (3.4% vs 2.7%; OR 1.27; 95% CI 0.88 to 1.82; P=.20).
  • A prespecified time-to-randomisation interaction at 3 months (P for homogeneity .03) suggested possible benefit when treatment began ≥24 hours after onset (OR 0.73; 95% CI 0.55 to 0.97), which remains hypothesis-generating.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • CATIS evaluates early inpatient BP lowering in non-thrombolysed ischaemic stroke; interpretation for reperfusion-treated populations should be anchored to reperfusion-specific BP thresholds and evidence.

Overall Takeaway

CATIS demonstrated that, in non-thrombolysed acute ischaemic stroke within 48 hours and SBP 140–<220 mm Hg, immediate protocolised BP reduction achieves robust physiological separation but does not improve death or functional outcome at either early (14 days/discharge) or standard (3 months) timepoints. The trial’s main value is in de-escalating routine early antihypertensive intensification in this population while sharpening questions about timing and context (hyperacute phase, reperfusion pathways, and severe stroke phenotypes).

Overall Summary

  • Early modest BP lowering (10%–25% in 24 h) after non-reperfused acute ischaemic stroke did not improve death or disability despite clear BP separation.
  • Temporary withholding of chronic antihypertensives during early hospitalisation was not associated with worse functional outcomes in this cohort.
  • Guideline recommendations to avoid routine early BP lowering (unless severely elevated or with compelling indications) align with CATIS findings in this setting.

Bibliography