Publication

• Title: Mechanical Thrombectomy in Ischemic Stroke With a Medium or Distal Arterial Occlusion: The DISCOUNT Randomized Clinical Trial
• Acronym: DISCOUNT
• Year: 2026
• Journal published in: JAMA
• Citation: Clarençon F, Bala F, Baptiste A, et al; DISCOUNT Investigators. Mechanical thrombectomy in ischemic stroke with a medium or distal arterial occlusion: the DISCOUNT randomized clinical trial. JAMA. Published online July 22, 2026.

Context & Rationale

• Background

  • Mechanical thrombectomy transformed the treatment of acute ischaemic stroke caused by proximal large-vessel occlusion, with the pivotal trials reporting absolute increases of approximately 11% to 31% in good functional outcome.
  • Those trials included few patients with more distal occlusions, and their results could not safely be extrapolated to smaller, more tortuous and more mobile arteries.
  • Primary medium or distal vessel occlusions account for up to one-fifth of acute ischaemic strokes. Although their median neurological severity is lower than that of proximal occlusions, injury to eloquent cortex may produce disabling aphasia, visual loss, neglect or limb dysfunction.
  • Intravenous thrombolysis achieves recanalisation in approximately 40% to 50% of these occlusions, creating a stronger medical comparator than in many proximal large-vessel occlusions.
  • Conversely, navigation and clot retrieval in small vessels may cause perforation, subarachnoid haemorrhage, branch avulsion or embolisation. The relevant question was therefore not whether thrombectomy could open the artery, but whether its net clinical effect remained favourable.
  • Observational series were conflicting and highly vulnerable to selection bias. While DISCOUNT was in progress, DISTAL and ESCAPE-MeVO also found no improvement in 90-day functional outcome with routine endovascular treatment in broadly selected medium or distal occlusions.¹²
• Research Question/Hypothesis

  • In adults with acute ischaemic stroke caused by a primary, isolated medium or distal intracranial arterial occlusion, does mechanical thrombectomy plus best medical treatment increase the proportion achieving a modified Rankin Scale score of 0–2 at 3 months compared with best medical treatment alone?
  • The trial was designed around the hypothesis that thrombectomy would produce a 13-percentage-point absolute increase in good clinical outcome, from 40% to 53%.
• Why This Matters

  • Thrombectomy was already being used for selected distal occlusions despite the absence of randomised evidence, exposing patients to an invasive procedure and health systems to substantial resource use.
  • A clinically important benefit would have extended a highly effective treatment to a large additional stroke population.
  • Failure to improve disability, particularly if accompanied by procedural haemorrhage, would argue against routine expansion and protect patients from a treatment whose angiographic success might not translate into recovery.

Design & Methods

• Research Question: Does mechanical thrombectomy, added to best medical treatment, improve 3-month functional independence in adults with acute ischaemic stroke caused by a primary and isolated medium or distal vessel occlusion?
• Study Type: Multicentre, French, open-label, parallel-group superiority randomised clinical trial with blinded endpoint assessment, conducted in 22 comprehensive stroke centres. Recruitment ran from November 2021 to November 2024, with follow-up completed in April 2025. Allocation was centralised through CLEANWEB in a 1:1 ratio. A minimisation algorithm balanced centre, age (<80 vs ≥80 years), NIHSS score (<10 vs ≥10), occlusion territory and planned intravenous thrombolysis; the first 10 allocations were simple randomisations and subsequent allocation retained a 10% random component.
• Population:
  • Adults aged ≥18 years with NIHSS ≥5 or significant aphasia potentially responsible for severe functional disability.
  • Presentation within 8 hours of symptom onset, or within 24 hours of last seen well when no hyperintense signal was present on FLAIR imaging.
  • Primary, isolated occlusion of distal M2 above the mid-height of the insula, M3, A1–A3 or P1–P3 on CT angiography or MR angiography.
  • Patients were functionally independent before the stroke: the principal eligibility threshold was prestroke mRS ≤1.
  • Important exclusions included tandem or secondary distal occlusion, two or more concomitant distal occlusions, concomitant proximal and distal occlusions, pre-existing haemorrhage in the supplied territory, proximal stenosis or stent that could impede retrieval, severe renal impairment, aortic or carotid dissection and anticipated limitations to anaesthesia.
  • The final time-window criteria were broader than the published protocol, which had originally specified expected groin puncture within 6 hours.³
• Intervention:
  • Mechanical thrombectomy plus best medical treatment, initiated as soon as possible and within 2 hours of randomisation.
  • Operators could use a stent retriever designed for medium or distal vessels, an aspiration catheter, or both, according to anatomy and technical conditions.
  • General anaesthesia, conscious sedation or local anaesthesia was permitted according to patient status and local practice.
  • Thrombectomy could be withheld after substantial neurological improvement following intravenous thrombolysis.
  • Non-contrast brain imaging was performed after thrombectomy and within 24 hours to identify complications.
• Comparison:
  • Best medical treatment alone.
  • Both groups could receive intravenous thrombolysis when eligible, within 4.5 hours or within 9 hours using EXTEND imaging criteria; the thrombolytic drug was selected by the treating neurologist.
  • Subsequent antiplatelet or anticoagulant treatment was determined by stroke aetiology and usual clinical care.
• Blinding: The procedural assignment could not be blinded. The 3-month mRS was assessed by an independent certified assessor blinded to allocation, and 24–48-hour recanalisation was assessed by a blinded core laboratory. Reperfusion immediately after thrombectomy was assessed locally and was therefore open to site-assessment bias.
• Statistics: A total of 488 patients was required to detect a 13-percentage-point absolute increase in mRS 0–2, from 40% to 53%, with 80% power, a two-sided 5% significance level, allowance for 5% loss to follow-up and one interim analysis. The prespecified interim analysis used Lan-DeMets/O’Brien-Fleming monitoring, with stopping for futility if conditional power was <30%. The primary analysis used a mixed logistic regression model in the full analysis set, adjusted for the minimisation factors with centre as a random effect. Missing 3-month mRS values among survivors were imputed by chained equations; per-protocol and prestroke-mRS sensitivity analyses were prespecified. Safety analyses were as treated, and secondary and subgroup analyses were not adjusted for multiplicity.
• Follow-Up Period: Clinical follow-up was to 3 months; vascular imaging for recanalisation was planned within 24–48 hours.

Key Results

This trial was stopped early after the first of two planned analyses, an interim analysis conducted after the primary endpoint had been evaluated in the first 163 randomised patients. On 8 December 2024, the data and safety monitoring board recommended stopping because efficacy was numerically lower with thrombectomy, symptomatic intracranial haemorrhage was overrepresented, and conditional power was <10%. By closure, 261 patients had been randomised and 244 comprised the full analysis set, compared with the planned sample of 488.

• Thrombectomy produced greater arterial recanalisation, but this 11.5-percentage-point angiographic advantage did not translate into a better primary functional outcome; the adjusted absolute difference in mRS 0–2 was −6.8 percentage points.
• Exploratory subgroup analyses showed no statistically significant treatment-effect heterogeneity:
  • Intravenous thrombolysis initiated/planned: OR 0.91; 95% CI 0.48 to 1.74; not planned: OR 0.57; 95% CI 0.22 to 1.47; interaction P=0.35.
  • NIHSS <10: OR 0.65; 95% CI 0.29 to 1.44; NIHSS ≥10: OR 0.85; 95% CI 0.35 to 2.06; interaction P=0.61.
  • Age <80 years: OR 0.58; 95% CI 0.27 to 1.24; age ≥80 years: OR 1.02; 95% CI 0.37 to 2.82; interaction P=0.36.
  • Distal MCA/ACA: OR 0.68; 95% CI 0.36 to 1.30; PCA: OR 1.40; 95% CI 0.24 to 7.96; interaction P=0.51.
• The principal safety signal was haemorrhagic: symptomatic intracranial haemorrhage occurred in 11% vs 3%, and any subarachnoid haemorrhage in 13% vs 2%, among patients analysed according to treatment actually received.

Internal Validity

• Randomisation and Allocation: Central web-based allocation and minimisation across major prognostic variables made allocation concealment credible and reduced selection bias after enrolment. Retaining a random component in the minimisation process prevented deterministic prediction.
• Post-randomisation Exclusions and Missing Data: Of 261 randomised patients, 17 were excluded from the full analysis set: 13 because pursuit consent was absent or use of data was opposed, and 4 because the trial indication was violated. The analysis was therefore not a strict intention-to-treat analysis. At 3 months, 217/244 (89%) had completed follow-up; the primary endpoint was unavailable in 7 thrombectomy-assigned and 2 control patients and was addressed by multiple imputation.
• Performance and Detection Bias: Treatment was necessarily open-label, allowing knowledge of allocation to influence peri-procedural care and decisions to proceed. The primary mRS assessment was independently blinded, and recanalisation imaging was core-laboratory assessed, materially limiting detection bias for the principal outcomes. Immediate procedural reperfusion remained site-reported.
• Protocol Adherence and Crossover: Only 100/123 patients (81%) allocated to thrombectomy received it. Nineteen improved neurologically, 2 had inaccessible occlusions and 2 had thrombus migration. No control patient underwent thrombectomy. This incomplete delivery diluted treatment separation in the primary analysis, but the per-protocol estimate remained directionally consistent and did not reveal benefit.
• Baseline Characteristics: Age, sex, NIHSS score, imaging modality, occlusion site and intravenous thrombolysis use were similar. However, hypertension was more common with thrombectomy (71% vs 51%), as was diabetes (24% vs 12%), whereas current smoking was more common in controls (12% vs 23%). These variables were not minimisation factors and were not included in the primary adjustment model.
• Heterogeneity: The trial combined distal M2, M3, ACA and PCA occlusions, which differ in vessel calibre, collateral supply, eloquence and natural history. It also permitted multiple devices, techniques and anaesthetic strategies. Territory was a minimisation factor, but small subgroup sizes left the study unable to exclude clinically important phenotype-specific effects.
• Timing: Median onset-to-randomisation was 3.5 hours in the thrombectomy arm and 3.3 hours in controls. Among treated patients, median onset-to-puncture time was 4.3 hours (IQR 3.3–5.6), and randomisation-to-puncture time was 0.7 hours (IQR 0.4–1.0). Delivery therefore met the procedural target, although the biological value of recanalisation may decline rapidly in small territories.
• Dose and Technical Delivery: The median number of passes was 1.5 (IQR 1–2). First-line treatment among the 100 treated patients was a stent retriever in 53%, aspiration in 33% and a combined approach in 5%; the device was not reported in 2% and no device was used in 7%. Rescue treatment was used in 19%. High procedural reperfusion suggests that technical failure alone does not explain the clinical result, but the heterogeneous device strategy prevents identification of an optimal technique.
• Separation of the Variable of Interest: Thrombectomy was delivered to 81% of the intervention group and 0% of controls. At 24–48 hours, recanalisation occurred in 85% vs 73%, an absolute difference of 11.5 percentage points. Thus, procedural separation was substantial but incomplete, while biological separation was real but modest.
• Adjunctive Therapy Use: Intravenous thrombolysis was initiated or planned in 66% vs 67%, providing excellent balance and a contemporary medical-treatment comparator rather than an undertreated control arm.
• Outcome Assessment: The blinded 3-month mRS is clinically meaningful and standard in stroke trials. However, dichotomising it at 0–2 sacrifices ordinal information and may be insensitive to focal but disabling deficits that are particularly relevant to medium or distal occlusions.
• Statistical Rigour: The sequential design, stopping rule, adjusted mixed model, bootstrap absolute difference, multiple imputation and sensitivity analyses were prespecified and appropriate. Precision was nevertheless reduced by stopping at approximately half the intended sample. Secondary and subgroup analyses were unadjusted for multiplicity. Safety analyses were as treated, which is clinically intuitive for procedural harms but breaks randomisation by placing the 23 untreated thrombectomy-assigned patients—many of whom improved early—in the no-thrombectomy group.

Conclusion on Internal Validity: Internal validity is moderate. Concealed randomisation, blinded primary outcome assessment, a strong contemporary control and concordant sensitivity analyses support the main efficacy conclusion, but early stopping, post-randomisation exclusions, incomplete intervention delivery, baseline imbalances and non-randomised safety comparisons limit precision and causal confidence around the magnitude of harm.

External Validity

• Population Representativeness: Participants were older (median 75 years), usually independent before stroke, and had predominantly mild-to-moderate deficits (median NIHSS 8; only 39% had NIHSS ≥10). Intravenous thrombolysis was planned in two-thirds. This closely represents an older, thrombolysis-eligible European MDVO population but not necessarily younger patients, severe deficits or thrombolysis-ineligible patients.
• Anatomical Applicability: Three-quarters had distal MCA occlusions, 19% PCA occlusions and 7% ACA occlusions. Proximal M2 occlusions, secondary emboli arising during proximal thrombectomy, multiple occlusions, tandem lesions and combined proximal/distal disease were excluded. The result should not be extended automatically to dominant proximal M2 occlusion.
• Selection in Practice: Clinicians may have treated patients perceived to have the greatest potential benefit outside the trial, particularly younger patients with severe or highly disabling deficits. Such pre-randomisation selection can narrow the enrolled phenotype even when formal eligibility is broad.
• Healthcare Setting: The study was conducted in 22 experienced French comprehensive stroke centres, with MRI used in approximately 80% and rapid access to specialist neurointervention. Results are most applicable to similarly organised systems; procedural risk could be greater in low-volume or resource-limited centres.
• Technical Generalisability: Operator discretion over stent retrievers, aspiration, anaesthesia and arterial access improves pragmatic relevance. However, the prespecified device menu included several devices not indicated for this use in their instructions for use, and named device exposure was not reported at patient level. Newer, purpose-built technologies could alter the balance, but that remains unproven.

Conclusion on External Validity: The findings are directly generalisable to functionally independent, older patients with primary isolated distal M2/M3, ACA or PCA occlusion, moderate deficits and high access to thrombolysis in comprehensive European stroke systems. Generalisability is limited for dominant or proximal M2 occlusion, severe deficits, low thrombolysis use, multiple occlusions, pre-existing disability and less experienced centres.

Strengths & Limitations

• Strengths:
  • Randomised, multicentre design across 22 comprehensive stroke centres.
  • Central concealed allocation with balancing of major prognostic factors.
  • Independent blinded assessment of the patient-centred primary endpoint.
  • Blinded core-laboratory assessment of follow-up recanalisation.
  • Contemporary best medical treatment, with closely balanced intravenous thrombolysis use.
  • Prespecified sequential monitoring and an independent data and safety monitoring board, allowing timely cessation when benefit became unlikely and harm emerged.
  • Detailed reporting of clinically important procedural harms, not merely technical success.
  • Pragmatic use of different thrombectomy strategies, reflecting specialist practice rather than a single proprietary device.
• Limitations:
  • The trial stopped with 244 analysed patients rather than 488, leaving wide confidence intervals and reduced power to detect a modest treatment effect.
  • The assumed 13-percentage-point benefit was optimistic, while the control event rate was far higher than the anticipated 40%: 68% achieved mRS 0–2.
  • Post-randomisation consent and eligibility exclusions prevented a strict intention-to-treat analysis.
  • Nineteen per cent of the intervention group did not receive thrombectomy, principally because of early neurological improvement.
  • Baseline hypertension and diabetes were appreciably more frequent in the intervention group.
  • Occlusion location at entry was not adjudicated by a central core laboratory, leaving scope for misclassification, especially around the distal-M2 boundary.
  • Forty-nine patients lacked adequate 24–48-hour imaging; immediate reperfusion was assessed by treatment sites rather than centrally.
  • The Results narrative states that 154 patients had adequate follow-up imaging, whereas the tabulated denominators are 98 and 97 and the footnote documents 49 missing scans, both implying 195. This appears to be a reporting inconsistency.
  • As-treated safety analyses did not preserve randomisation and may be confounded by the favourable prognosis of intervention-assigned patients whose thrombectomy was withheld after improvement.
  • The binary mRS 0–2 endpoint may obscure clinically meaningful differences across the full disability scale and subtle cortical deficits.
  • Anatomical, technical and device heterogeneity, together with small subgroups, limits identification of a responsive phenotype or a safer technique.
  • The prespecified health-economic evaluation was not reported in the primary publication.
  • Device manufacturers provided financial support, free devices or discounted devices, and several investigators disclosed relevant industry relationships. The funders were reported to have had no role in design, conduct, analysis or publication.

Interpretation & Why It Matters

• Routine thrombectomy is not supported

  For patients resembling those enrolled in DISCOUNT, adding thrombectomy did not improve 3-month functional independence and increased intracranial and subarachnoid haemorrhage. The trial therefore argues against treating the label “medium or distal vessel occlusion” as an automatic indication for intervention.
• Recanalisation is not a patient outcome

  The intervention opened more arteries—85% vs 73% at 24–48 hours—but did not improve the primary outcome. Procedural injury, reperfusion of tissue already destined for infarction, spontaneous or thrombolytic recanalisation in controls, and limited tissue-at-risk may all separate angiographic success from clinical benefit.
• The quality of the control arm matters

  Two-thirds of patients received or were planned to receive intravenous thrombolysis, and 68% of controls achieved mRS 0–2. Thrombectomy was therefore tested against effective modern stroke care, not therapeutic inactivity.
• The trial narrows rather than ends the question

  DISCOUNT provides evidence against broad routine treatment, but it does not exclude benefit in carefully selected patients with severe disabling deficits, dominant/proximal M2 occlusion, poor response or ineligibility for thrombolysis, favourable tissue imaging, or treatment with future safer devices.
• Clinical practice

  Outside a trial, the default for distal M2/M3, ACA and PCA occlusions matching DISCOUNT should be best medical treatment. Any decision to intervene should be exceptional, anatomy- and deficit-specific, made by an experienced multidisciplinary stroke team and explicit about uncertainty and haemorrhagic risk.

Controversies & Other Evidence

• Protocol evolution broadened the biological population: The published protocol required expected groin puncture within 6 hours, whereas the final trial allowed treatment within 8 hours and selected FLAIR-negative patients up to 24 hours.³ The amendment was transparently reported, but it increased temporal heterogeneity and makes it harder to determine whether very early treatment behaves differently.
• The primary endpoint creates a counterintuitive distribution: Thrombectomy produced more mRS 0–1 outcomes (41% vs 36%) but fewer mRS 0–2 outcomes (62% vs 68%). The published distribution shows that controls had more mRS 2 outcomes (32% vs 21%), while thrombectomy had more mRS 4 outcomes (15% vs 5%). This does not indicate hidden benefit; it demonstrates how a single dichotomy can discard clinically important information across the disability scale.
• Early stopping strengthens the qualitative message but weakens quantitative precision: Conditional power below 10%, a point estimate favouring medical treatment and excess haemorrhage made continuation difficult to justify. Nevertheless, the adjusted absolute-difference interval (−19.4 to 5.7 percentage points) still permits modest benefit or substantial harm; it excludes the prespecified 13-point benefit but does not provide a precise estimate of the true effect.
• Safety is compelling but not a fully randomised comparison: The as-treated analysis compares 100 patients who underwent thrombectomy with 144 who did not, including 23 intervention-assigned patients whose procedure was withheld, mostly after early improvement. This may magnify between-group prognostic differences. The concordant excess of perforation, embolisation, subarachnoid haemorrhage and symptomatic intracranial haemorrhage nevertheless forms a coherent procedural-harm signal.
• DISCOUNT aligns with the first two definitive trials: DISTAL found no improvement in 90-day ordinal mRS (common OR 0.90; 95% CI 0.67 to 1.22; P=0.50), with symptomatic intracranial haemorrhage in 5.9% vs 2.6%.¹ ESCAPE-MeVO found mRS 0–1 in 41.6% vs 43.1% (adjusted rate ratio 0.95; 95% CI 0.79 to 1.15; P=0.61), with mortality 13.3% vs 8.4% (adjusted hazard ratio 1.82; 95% CI 1.06 to 3.12) and symptomatic intracranial haemorrhage 5.4% vs 2.2%.² The accompanying editorial concluded that thrombectomy could not be assumed to remain beneficial merely by moving distally along the vascular tree.⁴
• Pooled evidence reinforced the haemorrhagic concern: A 2025 meta-analysis of the first three randomised trials, using the then-available DISCOUNT data, found no improvement in excellent outcome (OR 0.92; 95% CI 0.72 to 1.17), good outcome (OR 0.87; 95% CI 0.70 to 1.09) or favourable outcome (OR 0.84; 95% CI 0.64 to 1.10), but more symptomatic intracranial haemorrhage (OR 2.18; 95% CI 1.24 to 3.83).⁵
• Selection bias may have removed the patients most likely to benefit: Younger patients with severe or highly eloquent deficits may have been treated outside randomisation because clinicians perceived equipoise to be absent. Contemporary commentary has emphasised that medium and distal occlusion is not one disease entity and that anatomy, tissue at risk and deficit severity must be considered jointly.⁶
• ORIENTAL-MeVO showed that a more selected population can behave differently: In 563 Chinese patients with moderate-to-severe deficits, median NIHSS 10, median age 71 years and intravenous thrombolysis use of 36.6%, thrombectomy increased mRS 0–2 at 90 days from 46.6% to 58.6% (adjusted rate ratio 1.24; 95% CI 1.07 to 1.44; P=0.004), while symptomatic intracranial haemorrhage occurred in 4.7% vs 2.2%.⁷ Its accompanying editorial interpreted this as a narrowing of the target population rather than a reversal of the earlier trials.⁸ Differences in deficit severity, imaging selection, age, thrombolysis use, anatomy and workflow make direct reconciliation uncertain.
• There was no late functional rescue in DISTAL: At 12 months, endovascular treatment did not improve the ordinal mRS (adjusted common OR 0.81; 95% CI 0.59 to 1.12; P=0.20) or overall survival (hazard ratio 1.46; 95% CI 0.93 to 2.30; P=0.10), supporting the durability of the 90-day result in mild-to-moderate MDVO.⁹
• The emerging consensus is phenotype-specific treatment: A 2026 review framed the four major trials as evidence that vessel labels alone are inadequate and advocated selection by anatomy, deficit severity, tissue viability and procedural feasibility.¹⁰ The 2026 Society of Vascular and Interventional Neurology focused guideline concluded that routine thrombectomy for distal medium-vessel occlusion was not supported, while considering treatment reasonable for disabling acute dominant M2 occlusion.¹¹ This guideline predated publication of the final DISCOUNT manuscript and therefore incorporated its conference-level rather than final peer-reviewed data.
• Future evidence needs individual-patient rather than trial-level synthesis: Aggregate averages mix dominant and non-dominant M2, M3, ACA and PCA occlusions; mild and severe deficits; thrombolysis-responsive and thrombolysis-ineligible patients; and different devices. An individual-patient-data meta-analysis with prespecified anatomy, severity, imaging, timing and thrombolysis interactions is more likely than another broadly inclusive trial to identify a credible beneficial subgroup.

Summary

• DISCOUNT randomised adults with primary isolated distal M2/M3, ACA or PCA occlusion to thrombectomy plus medical treatment or medical treatment alone across 22 French stroke centres.
• The trial stopped at its first interim analysis, after evaluation of the first 163 participants, for futility, lower observed efficacy and excess symptomatic intracranial haemorrhage; 244 patients were included in the full analysis set rather than the planned 488.
• Good functional outcome at 3 months occurred in 62% vs 68% (adjusted OR 0.73; 95% CI 0.40 to 1.31; P=0.29; adjusted absolute difference −6.8 percentage points).
• Thrombectomy increased 24–48-hour recanalisation from 73% to 85%, but symptomatic intracranial haemorrhage increased from 3% to 11% and subarachnoid haemorrhage from 2% to 13% in as-treated analyses.
• The trial provides strong evidence against routine thrombectomy for DISCOUNT-like medium or distal occlusions, while leaving uncertainty for carefully selected severe, dominant-M2, thrombolysis-ineligible or imaging-favourable phenotypes.

Overall Takeaway

DISCOUNT is an important practice-defining trial because it demonstrates that successful recanalisation in smaller distal arteries is not sufficient to justify an invasive procedure: functional benefit was absent and haemorrhagic harm increased. Together with DISTAL and ESCAPE-MeVO, it moves clinical practice away from routine vessel-based thrombectomy and towards cautious, phenotype-specific selection informed by deficit severity, anatomy, imaging, thrombolysis response and procedural risk.

Bibliography

• 1.Psychogios MN, Brehm A, Ribo M, et al; DISTAL Investigators. Endovascular treatment for stroke due to occlusion of medium or distal vessels. N Engl J Med. 2025;392(14):1374-1384.
• 2.Goyal M, Ospel JM, Ganesh A, et al; ESCAPE-MeVO Investigators. Endovascular treatment of stroke due to medium-vessel occlusion. N Engl J Med. 2025;392(14):1385-1395.
• 3.Clarençon F, Durand-Zaleski I, Premat K, et al. Evaluation of mechanical thrombectomy in acute ischemic stroke related to a distal arterial occlusion: a randomized controlled trial. Int J Stroke. 2024;19(3):367-372.
• 4.Mocco J. Medium- and distal-vessel occlusion—the limit of thrombectomy? N Engl J Med. 2025;392(14):1440-1442.
• 5.Jhou HJ, Wang WS, Lee CH, Yang LY, Chen PH. Thrombectomy for distal medium vessel occlusion: a meta-analysis of randomized controlled trials. J Am Heart Assoc. 2025;14(20):e042299.
• 6.Murillo-Olaizola A, Rodrigo-Gisbert M, Requena M, et al. Endovascular treatment for medium and distal vessel occlusion stroke: are we missing the point? Stroke Vasc Interv Neurol. 2026;6(2):e002137.
• 7.Hu W, Jing X, Chen Z, et al. Endovascular treatment of medium-vessel-occlusion strokes. N Engl J Med. 2026;394(19):1894-1904.
• 8.Ospel JM, Hill MD. Endovascular therapy for medium-vessel occlusion stroke—narrowing the target population. N Engl J Med. 2026;394(19):1955-1957.
• 9.Fischer U, Brehm A, Ribo M, et al; DISTAL Investigators. Endovascular treatment for medium or distal vessel occlusion stroke (DISTAL): 12-month outcomes of a multicentre, open-label, randomised trial. Lancet Neurol. 2026;25(6):571-580.
• 10.Goyal M, Lun R, Ospel JM, Ganesh A. Navigating the complexity and heterogeneity of thrombectomy decision-making for medium-vessel occlusion stroke: where do we go from here? Stroke Vasc Interv Neurol. 2026;6(4):e002439.
• 11.Majidi S, Mehta A, Al-Kasab S, et al. Endovascular thrombectomy in medium and distal vessel occlusions: a focused guideline from the Society of Vascular and Interventional Neurology Guidelines and Practice Standards Committee. Stroke Vasc Interv Neurol. 2026;6(4):e002314.