Publication

• Title: The International Stroke Trial (IST): a randomised trial of aspirin, subcutaneous heparin, both, or neither among 19 435 patients with acute ischaemic stroke
• Acronym: IST
• Year: 1997
• Journal published in: The Lancet
• Citation: International Stroke Trial Collaborative Group. The International Stroke Trial (IST): a randomised trial of aspirin, subcutaneous heparin, both, or neither among 19 435 patients with acute ischaemic stroke. Lancet. 1997;349:1569–81.

Context & Rationale

• Background

  • Immediate antithrombotic therapy after suspected ischaemic stroke was common, but practice varied widely (aspirin, subcutaneous heparin, both, or neither).
  • Aspirin was established for secondary prevention, but its net benefit–harm balance in the first days after stroke was uncertain (recurrent ischaemia prevention vs intracranial/extracranial bleeding).
  • Heparin was used with goals spanning recurrent embolism prevention, stroke progression prevention, and venous thromboembolism reduction, but with substantial uncertainty about intracranial haemorrhage risk and overall net clinical benefit.
  • Existing trials and observational experience were insufficiently powered and heterogeneous, leaving clinically important modest absolute effects unresolved.
  • A very large pragmatic trial was needed to detect or refute small absolute differences that could still be important at a population level.
• Research Question/Hypothesis

  • In hospitalised patients with suspected acute ischaemic stroke within 48 hours, does early aspirin (300 mg daily) and/or subcutaneous unfractionated heparin (5,000 IU twice daily or 12,500 IU twice daily), given for up to 14 days, reduce (a) death within 14 days and (b) death or dependency at 6 months compared with avoiding these treatments?
  • Does the heparin dose materially shift the balance between preventing recurrent ischaemic events and causing intracranial/extracranial haemorrhage?
• Why This Matters

  • Because stroke incidence is high, small absolute effects (a few events per 1,000 treated) can translate into substantial public health impact.
  • Immediate post-stroke antithrombotic decisions are time-critical and often precede definitive aetiological classification; robust evidence was needed to standardise care.
  • IST directly informed modern recommendations favouring early aspirin (once haemorrhage is excluded) and avoiding routine therapeutic anticoagulation in unselected acute ischaemic stroke.

Design & Methods

• Research Question: Whether early aspirin and/or subcutaneous unfractionated heparin, started within 48 hours of acute ischaemic stroke, improve early survival and longer-term functional outcome, and how benefit–harm varies by heparin dose.
• Study Type: Pragmatic, randomised, multicentre, international, investigator-initiated collaborative trial; open-label (no placebo); partial factorial allocation with central telephone randomisation using minimisation/stratification by prognostic variables; hospital-based acute stroke care across 467 hospitals in 36 countries.
• Population:
  • Setting: in-patient acute stroke management (stroke units and general wards) across diverse health systems.
  • Inclusion: clinical diagnosis of acute stroke judged likely ischaemic; symptom onset ≤48 hours; treating clinician was uncertain whether aspirin and/or heparin should be given (the “uncertainty principle”).
  • Imaging: CT before randomisation was not mandatory; if CT was available and showed intracranial haemorrhage, the patient was not eligible for randomisation to aspirin/heparin for presumed ischaemic stroke.
  • Key exclusions (pragmatic): clear indication for, or contraindication to, aspirin or heparin (including major bleeding risk); and (for heparin randomisation) patients already receiving long-term oral anticoagulants.
• Intervention:
  • Aspirin allocation: 300 mg once daily, started immediately after randomisation and continued for up to 14 days (or until discharge).
  • Heparin allocation: subcutaneous unfractionated heparin started immediately after randomisation and continued for up to 14 days (or until discharge), at either 5,000 IU twice daily (low-dose) or 12,500 IU twice daily (medium-dose).
  • Factorial structure: where clinicians were uncertain about both treatments, patients could be allocated to aspirin vs avoid aspirin and to heparin (low or medium dose) vs avoid heparin; where uncertainty applied to only one treatment, partial factorial randomisation was used for that treatment.
• Comparison:
  • Avoid aspirin: no routine antiplatelet during the trial treatment period unless a clear indication emerged.
  • Avoid heparin: no routine subcutaneous heparin during the trial treatment period unless a clear indication emerged.
  • Co-interventions: other stroke care elements were not protocolised; treating clinicians could introduce open-label antithrombotics if clinical circumstances changed, and these were recorded.
• Blinding: None (open-label; no placebo). Implication: robust for mortality, but in-hospital detection/reporting of haemorrhagic outcomes could be influenced by lack of blinding.
• Statistics: Planned enrolment ≥20,000 to minimise the risk of a false-negative result for a moderate treatment effect (targeting approximately a 15% proportional reduction in 14-day mortality from an expected ~10% baseline); formal alpha and beta/power were not explicitly reported. Primary analyses were intention-to-treat by allocated treatment (aspirin vs avoid aspirin; heparin vs avoid heparin), with additional comparisons of heparin dose; adjusted analyses used a pre-specified prognostic model (O–E and V method) incorporating baseline prognostic factors.
• Follow-Up Period: Early outcomes to 14 days (or discharge if earlier), plus vital status and dependency/recovery at 6 months using a structured follow-up questionnaire (with telephone follow-up when needed).

Key Results

This trial was not stopped early. Interim analyses were reviewed about once yearly by an independent data monitoring committee, and recruitment proceeded to completion.

• Aspirin allocation produced a modest absolute reduction in early recurrent ischaemic stroke (2.8% vs 3.9%) but increased major extracranial bleeding (1.1% vs 0.6%); longer-term death or dependency showed a small benefit only after prognostic adjustment.
• Heparin allocation reduced recurrent ischaemic stroke (2.9% vs 3.8%) and pulmonary embolism (0.5% vs 0.8%) but substantially increased haemorrhagic stroke (1.2% vs 0.4%) and major extracranial bleeding (1.3% vs 0.4%), yielding no improvement in death or dependency at 6 months.
• Heparin dose mattered: medium-dose heparin (12,500 IU bd) increased transfused or fatal extracranial haemorrhage (2.0% vs 0.6%) and haemorrhagic stroke (1.8% vs 0.7%) compared with low-dose heparin (5,000 IU bd), and increased death or non-fatal stroke within 14 days (12.6% vs 10.8%).

Internal Validity

• Randomisation and Allocation:
  • Central telephone randomisation with minimisation/stratification based on a pre-specified prognostic model supports allocation concealment and baseline balance.
  • Pragmatic “uncertainty principle” eligibility reduces selection by preference but embeds clinician judgement into enrolment (a design trade-off rather than a flaw per se).
• Drop out or exclusions:
  • No post-randomisation withdrawals by design; participants could deviate from assigned therapy but remained in follow-up.
  • Follow-up completeness was very high: 14-day outcome data were essentially complete (≥99.9% reported), and 6-month outcome data were available for ~99.2% of participants in each main comparison.
• Performance/Detection Bias:
  • Open-label delivery (no placebo) introduces risk of differential co-interventions and differential detection/reporting of bleeding outcomes in hospital.
  • Mortality outcomes are objective; functional outcome at 6 months used a standardised questionnaire with central follow-up processes, mitigating (but not eliminating) detection bias.
• Protocol Adherence:
  • Heparin allocation: 89% received heparin for the full scheduled period (low-dose 90%; medium-dose 88%); avoid heparin: 94% received no heparin during the scheduled period.
  • Aspirin allocation: 92% received aspirin for the full scheduled period; avoid aspirin: 93% received no antiplatelet during the scheduled period.
  • Mean scheduled treatment duration was 11 days.
• Baseline Characteristics:
  • Groups were well balanced by minimisation; overall cohort included 61% aged >70 years, 23% with impaired consciousness at baseline, and 16% with atrial fibrillation.
• Heterogeneity:
  • Clinical heterogeneity was intentionally high (467 hospitals, 36 countries; broad “suspected ischaemic stroke” entry), which increases real-world relevance but can dilute treatment effects if benefits are subtype- or time-dependent.
• Timing:
  • Median time from symptom onset to randomisation was 19 hours (15% within 6 hours; 32% 7–12 hours; 34% 13–24 hours; 19% 25–48 hours), potentially attenuating benefits expected to be greatest earlier.
• Dose:
  • Aspirin dose (300 mg daily) reflects pragmatic early antiplatelet practice and is plausibly adequate for platelet inhibition in the acute phase.
  • Heparin dosing showed a clear dose–toxicity gradient: medium-dose increased haemorrhagic complications and worsened early composite outcomes compared with low-dose.
• Separation of the Variable of Interest:
  • Heparin: full-period exposure 89% (allocated heparin) vs 94% with no heparin (allocated avoid heparin).
  • Aspirin: full-period exposure 92% (allocated aspirin) vs 93% with no antiplatelet (allocated avoid aspirin).
  • CT timing (a key co-variable influencing misclassification of haemorrhage): CT done before randomisation in 67%, after randomisation in 29%, and never in 4%.
• Key Delivery Aspects:
  • Entry based on clinical diagnosis and uncertainty allowed enrolment before definitive imaging or aetiological classification; this maximised recruitment and generalisability but increased the likelihood of including some patients with primary haemorrhage or non-stroke mimics.
  • Partial factorial randomisation preserved pragmatism but requires careful interpretation: each main comparison (aspirin vs avoid; heparin vs avoid) pools across the other allocation strata.
• Outcome Assessment:
  • Early outcomes were ascertained during hospitalisation (to 14 days), with prespecified vascular and bleeding event categories.
  • Six-month dependency/recovery status was assessed using a structured questionnaire with central follow-up processes.
• Statistical Rigor:
  • Large sample size and near-complete follow-up reduce random error and attrition bias.
  • Use of pre-specified prognostic adjustment for 6-month functional outcome addresses baseline severity imbalances and improves precision; primary comparisons remained intention-to-treat by allocation.

Conclusion on Internal Validity: Overall, internal validity appears moderate to strong given concealed central randomisation, high adherence and near-complete follow-up; the principal limitation is the open-label design with variable pre-randomisation imaging, which could influence ascertainment of early haemorrhagic outcomes and introduces some diagnostic misclassification risk.

External Validity

• Population Representativeness:
  • Very broad inclusion across 36 countries and 467 hospitals supports applicability to routine hospital stroke care.
  • Pragmatic clinical diagnosis (with CT not required before randomisation) mirrors real-world decision-making in settings where imaging may be delayed or unavailable.
• Applicability:
  • Findings are most applicable to unselected acute stroke presentations within 48 hours, including later presenters (median 19 hours) and mixed suspected stroke mechanisms.
  • In modern systems with rapid imaging and reperfusion pathways, the core conclusions still translate: start antiplatelet therapy after haemorrhage exclusion; avoid routine full-dose anticoagulation in unselected acute ischaemic stroke.
  • Heparin dose findings remain relevant to balancing venous thromboembolism prevention against intracranial bleeding risk in immobilised stroke patients.

Conclusion on External Validity: Generalisability is high for routine hospital management of suspected acute ischaemic stroke, particularly where early definitive aetiological classification is not available; applicability to modern hyperacute reperfusion cohorts is directionally consistent but absolute effects may differ with earlier presentation and contemporary co-interventions.

Strengths & Limitations

• Strengths:
  • Very large sample size (19,435), enabling detection of modest absolute effects.
  • International multicentre pragmatic design supports broad generalisability.
  • Central randomisation with minimisation/stratification and near-complete follow-up.
  • Factorial structure enabled simultaneous evaluation of aspirin and heparin strategies and heparin dose.
• Limitations:
  • Open-label (no placebo) with potential for performance and detection bias, particularly for bleeding outcomes.
  • CT imaging not mandatory prior to randomisation (CT before randomisation 67%; after 29%; never 4%), increasing diagnostic heterogeneity and risk of haemorrhage misclassification at baseline.
  • Median randomisation at 19 hours may attenuate time-dependent benefits of early antithrombotic therapy.
  • Broad “uncertainty principle” inclusion yields clinical heterogeneity that can dilute mechanistically targeted treatment effects.

Interpretation & Why It Matters

• Immediate aspirin

  Produces a small but clinically meaningful reduction in early recurrent ischaemic stroke (2.8% vs 3.9%) at the cost of increased major extracranial haemorrhage (1.1% vs 0.6%); supports starting aspirin early after haemorrhage is excluded in suspected ischaemic stroke.
• Routine heparin

  Does not improve death or dependency at 6 months (62.9% vs 62.9%) despite reducing recurrent ischaemic stroke and pulmonary embolism, because it substantially increases haemorrhagic stroke and major extracranial bleeding; supports avoiding routine therapeutic-dose heparin in unselected acute ischaemic stroke.
• Dose–toxicity gradient

  Medium-dose heparin (12,500 IU bd) worsened early net outcomes compared with low-dose (5,000 IU bd), highlighting that “more anticoagulation” in acute stroke can convert potential benefit into net harm.

Controversies & Subsequent Evidence

• Pragmatic enrolment based on clinician uncertainty and absence of mandatory pre-randomisation CT was highlighted as both a strength (real-world relevance) and a limitation (diagnostic heterogeneity and potential baseline haemorrhage misclassification).¹
• Interpretation of the aspirin signal (small absolute benefit with bleeding trade-offs) and the emphasis placed on composite/adjusted outcomes prompted debate in the immediate post-publication period.²
• The CAST trial independently demonstrated modest absolute benefit from early aspirin in acute ischaemic stroke, supporting a consistent direction of effect across large pragmatic trials.³
• A combined analysis of IST and CAST (≈40,000 randomised patients) quantified the early aspirin benefit–harm profile at scale and reinforced that the net effects are small in absolute terms but important at population level.⁴
• Systematic review evidence has not supported routine anticoagulation for unselected acute ischaemic stroke because reductions in recurrent ischaemic events are offset by increased intracranial and extracranial bleeding, with no improvement in longer-term death or dependency.⁵
• Major contemporary guidelines incorporate these findings by recommending early aspirin after haemorrhage exclusion and advising against routine urgent anticoagulation in unselected acute ischaemic stroke (with anticoagulation reserved for specific indications and timing decisions made case-by-case).⁶

Summary

• IST randomised 19,435 patients with suspected acute ischaemic stroke within 48 hours to aspirin and/or subcutaneous heparin (low or medium dose) versus avoidance in a pragmatic international design.
• Aspirin reduced early recurrent ischaemic stroke (2.8% vs 3.9%) and improved an early composite endpoint (11.3% vs 12.4%), while increasing major extracranial bleeding (1.1% vs 0.6%).
• Heparin reduced recurrent ischaemic stroke (2.9% vs 3.8%) and pulmonary embolism (0.5% vs 0.8%) but markedly increased haemorrhagic stroke (1.2% vs 0.4%) and major extracranial bleeding (1.3% vs 0.4%).
• Neither aspirin nor heparin materially reduced death within 14 days; heparin did not improve death or dependency at 6 months (62.9% vs 62.9%).
• Heparin dose was crucial: medium-dose heparin worsened early net outcomes compared with low-dose, emphasising a clinically important dose–toxicity trade-off.

Further Reading

Other Trials

• 1997CAST (Chinese Acute Stroke Trial) Collaborative Group. CAST: randomised placebo-controlled trial of early aspirin use in 20,000 patients with acute ischaemic stroke. Lancet. 1997;349:1641–9.
• 1998TOAST Investigators. Low-molecular-weight heparinoid, ORG 10172 (danaparoid), and outcome after acute ischaemic stroke: a randomised controlled trial. JAMA. 1998;279:1265–72.
• 2000Berge E, Abdelnoor M, Nakstad PH, Sandset PM. Low molecular-weight heparin versus aspirin in patients with acute ischaemic stroke and atrial fibrillation: a double-blind randomised study. Lancet. 2000;355:1205–10.
• 2001Bath PMW, Lindenstrøm E, Boysen G, et al. Tinzaparin in acute ischaemic stroke trial (TAIST): a randomised aspirin-controlled trial. Lancet. 2001;357:702–10.
• 2007Wong KS, Chen C, Ng PW, et al. Low-molecular-weight heparin compared with aspirin for acute ischaemic stroke with large artery occlusive disease: a randomised trial. Lancet Neurol. 2007;6:615–22.

Systematic Review & Meta Analysis

• 2000Chen ZM, Sandercock P, Pan HC, et al. Indications for early aspirin use in acute ischaemic stroke: a combined analysis of 40,000 randomised patients from CAST and IST. Stroke. 2000;31:1240–9.
• 2021Sandercock PAG, Counsell C, Kamal AK. Anticoagulants for acute ischaemic stroke. Cochrane Database Syst Rev. 2021;10:CD000024.
• 2022Sandercock PAG, Counsell C, Tseng MC, Cecconi E. Oral antiplatelet therapy for acute ischaemic stroke. Cochrane Database Syst Rev. 2022;10:CD000029.
• 2012Wang QS, et al. Low-molecular-weight heparin versus aspirin in early management of acute ischaemic stroke: a systematic review and meta-analysis. Stroke. 2012.
• 2015Thompson BB, Bejot Y, Caso V, et al. Targeting aspirin in acute disabling ischaemic stroke: an individual patient data meta-analysis. Int J Stroke. 2015.

Observational Studies

• 2015Paciaroni M, Agnelli G, Falocci N, et al. Early recurrence and major bleeding in patients with acute ischaemic stroke and atrial fibrillation treated with anticoagulants: the RAF study. JAMA Neurol. 2015.
• 2021Meya L, et al. Oral anticoagulants in atrial fibrillation patients with ischaemic stroke and cerebral microbleeds. Stroke. 2021.
• 2022Kimura S, et al. Rule for starting direct oral anticoagulants after ischaemic stroke in patients with atrial fibrillation. Stroke. 2022.
• 2024Dehbi HM, et al. Optimal timing of anticoagulation after acute ischaemic stroke with atrial fibrillation: updated evidence synthesis. Lancet. 2024.

Guidelines

• 2019Powers WJ, Rabinstein AA, Ackerson T, et al. Guidelines for the early management of patients with acute ischaemic stroke: 2019 update. Stroke. 2019;50:e344–e418.
• 2021Kleindorfer DO, Towfighi A, Chaturvedi S, et al. Guideline for the prevention of stroke in patients with stroke and transient ischaemic attack. Stroke. 2021.
• 2020Hindricks G, Potpara T, Dagres N, et al. Guidelines for the diagnosis and management of atrial fibrillation. Eur Heart J. 2020.
• 2016European Stroke Organisation. Guidelines for prophylaxis for venous thromboembolism in immobile patients with acute ischaemic stroke. Eur Stroke J. 2016.
• 2024Heart and Stroke Foundation of Canada. Canadian Stroke Best Practice Recommendations: acute stroke management guideline update. Can J Neurol Sci. 2024.

Notes

• IST is methodologically pivotal as a demonstration that very large pragmatic trials can resolve modest absolute effects and quantify benefit–harm trade-offs for ubiquitous therapies.
• Key contextual detail: CT was performed before randomisation in 67%, after randomisation in 29%, and not at all in 4%, embedding real-world diagnostic constraints into the trial’s causal estimand.

Overall Takeaway

IST is a landmark because it quantified, at scale, that immediate aspirin confers a small absolute benefit on early recurrent ischaemic events with a measurable bleeding trade-off, while routine subcutaneous heparin in unselected acute ischaemic stroke yields no net long-term benefit due to increased intracranial and extracranial haemorrhage. Its pragmatic, international design under real-world diagnostic constraints shaped modern guideline recommendations and remains foundational for interpreting absolute risk trade-offs in acute stroke therapeutics.

Bibliography

• 1Bousser MG. Aspirin or heparin immediately after a stroke? Lancet. 1997;349:1564–5.
• 2Barer D. Interpretation of IST and CAST stroke trials. Lancet. 1997;350:440.
• 3CAST (Chinese Acute Stroke Trial) Collaborative Group. CAST: randomised placebo-controlled trial of early aspirin use in 20,000 patients with acute ischaemic stroke. Lancet. 1997;349:1641–9.
• 4Chen ZM, Sandercock P, Pan HC, et al. Indications for early aspirin use in acute ischaemic stroke: a combined analysis of 40,000 randomised patients from CAST and IST. Stroke. 2000;31:1240–9.
• 5Sandercock PAG, Counsell C, Kamal AK. Anticoagulants for acute ischaemic stroke. Cochrane Database Syst Rev. 2021;10:CD000024.
• 6Powers WJ, Rabinstein AA, Ackerson T, et al. Guidelines for the early management of patients with acute ischaemic stroke: 2019 update. Stroke. 2019;50:e344–e418.