CIP · development draft v0.1

AIRCHILL-STROKE Clinical Investigation Plan

Operational CIP for peri-EVT AIRCHILL in adults with anterior-circulation LVO who already require general anaesthesia and invasive ventilation. The protocol explicitly prohibits delaying arterial puncture or reperfusion for study treatment.

1. Administrative information

Sponsor: to be designated. Protocol: AIRCHILL-STROKE. Device: AIRCHILL investigational ventilation/cooling system. Version: development v0.1. Stages: ST-1 feasibility → ST-2 signal finding → ST-3 confirmatory.

2. Objective

Determine whether very early respiratory cooling delivered around thrombectomy improves 90-day functional outcome while preserving anaesthesia, haemodynamics and EVT speed/quality.

3. Design

Prospective multicentre randomised controlled open-treatment, blinded-endpoint investigation. Central 1:1 allocation before arterial puncture or at the earliest feasible peri-EVT point.

4. Population

Adults with acute anterior-circulation LVO selected for EVT who require general anaesthesia and invasive ventilation for clinical reasons independent of research participation.

Draft inclusion

• Age ≥18 years.
• Protocol-defined anterior-circulation LVO eligible for EVT.
• General anaesthesia/invasive ventilation clinically indicated.
• Imaging profile within final ASPECTS/core/penumbra criteria.
• Randomisation within the final early/extended time window.
• Permitted consent pathway available.

Draft exclusion

• Primary intracranial haemorrhage.
• Severe pre-stroke disability above the final locked threshold.
• Refractory haemodynamic or respiratory instability incompatible with study gas delivery.
• Terminal illness or clear non-survivable condition.
• Device-specific contraindication.
• Any circumstance in which study procedures would delay reperfusion.

5. Intervention

AIRCHILL begins after airway security and is integrated in parallel with anaesthesia and EVT preparation. Device setup must not extend door-to-puncture or puncture-to-reperfusion intervals. The final dose is defined by patient-side temperature, humidity, FiO₂, flow/minute ventilation, pressures and duration.

6. Comparator

Standard EVT, anaesthesia, ventilation and temperature management without active AIRCHILL respiratory cooling.

Ventilatory co-intervention control: PaCO₂/EtCO₂ targets, timing and protocol deviations must be prespecified and kept comparable between groups. Carbon dioxide is itself a potent cerebral vasodilator and is being prospectively manipulated during mechanically ventilated EVT in NCT05051397 (ClinicalTrials.gov record updated 3 September 2026). AIRCHILL therefore must not create or be confounded by systematic hypercapnia/normocapnia differences when estimating a cooling effect.

7. Schedule of assessments

Screening/imaging; pre-randomisation baseline; randomisation/device start; continuous peri-procedural device and physiological data; puncture; reperfusion; immediate post-EVT; 2–6 h; 24 h imaging/NIHSS; 48–72 h; discharge; 30 d; 90 d blinded mRS/EQ-5D. Shared detailed matrix: SoA / CRF.

8. Endpoints

Primary ST-3: blinded ordinal 90-day mRS shift. Key secondary: mortality, mRS 0–2 and 0–1, NIHSS, EQ-5D-5L, symptomatic ICH, malignant oedema/decompressive surgery, ICU/hospital stay. ST-2 imaging: final infarct volume, infarct growth, penumbral salvage where available, oedema metrics. Procedural: puncture-to-reperfusion, eTICI, passes, rescue treatment. Safety: device-related SAE, ventilation/gas-exchange compromise, haemodynamic events, arrhythmia, pneumonia and any study-attributable EVT delay.

9. Randomisation/blinding

Central concealed 1:1 randomisation, variable blocks, limited stratification by site and key baseline severity/time-window variables. Treating team unblinded; 90-day mRS assessor and imaging core laboratory blinded.

10. Sample size

Working ST-3 anchor: ~1,600 participants for common OR ~1.35. ~1,300 for OR ~1.40; >2,000 may be required for OR ~1.30. Final simulation uses ST-2 AIRCHILL mRS distributions, the final covariates/estimand, missingness and interim design.

11. Statistics

ITT primary analysis using proportional-odds ordinal logistic regression, with prespecified assumption checks and robust sensitivity analysis. Imaging analyses use blinded core-lab continuous endpoints with strategies preventing survivorship bias. See SAP.

12. DSMB and stopping

Independent DSMB reviews respiratory/device safety, haemodynamics, symptomatic ICH, mortality, malignant oedema and any treatment-attributable delay in puncture/reperfusion. Formal ST-3 efficacy/futility interim near 50% information is planned subject to final statistical design. See DSMB Charter.

13. Monitoring

Critical data include eligibility, baseline imaging, consent, randomisation, device configuration/start, arterial puncture/reperfusion timestamps, eTICI, 24-h imaging, SAE/device deficiency and 90-day mRS. Central monitoring flags treatment delays and site-specific workflow drift. See operations plan.

14. Ethics

No participant is intubated solely for AIRCHILL. Consent follows local prospective representative or emergency/deferred pathways, with participant consent if capacity returns.

15. US public-data site-selection layer

For US expansion, site selection now uses a staged public-data screen before protocol feasibility. CMS Medicare inpatient hospital records are used only as broad observed burden and payment signals; they are not interpreted as LVO, EVT or AIRCHILL-eligible counts. The first-pass stroke screen uses the relevant cerebrovascular MS-DRG family (including 061–066), then requires centre-specific confirmation of EVT volume, anterior-circulation LVO volume, clinically indicated general-anaesthesia/invasive-ventilation fraction, door-to-puncture workflow, competing trials and expected randomisations per centre-month. ClinicalTrials.gov activity is used as a research-capability signal, not as proof of AIRCHILL suitability. See US Clinical & Public Data Intelligence →.

Site-score framework: 20% CMS cerebrovascular signal; 20% EVT/LVO volume; 15% GA/invasive-ventilation eligible fraction; 15% active stroke-trial capability; 10% door-to-puncture workflow; 10% emergency-research/consent readiness; 10% expected randomisations per centre-month. Final weights remain development parameters until validated against run-in recruitment.

16. Health-economic data architecture

The trial distinguishes four layers: (A) observed CMS discharges/payments; (B) observed AIRCHILL trial resource use and outcomes; (C) prespecified outcome-improvement scenarios; and (D) long-term QALY/cost models. Historical CMS average payments are neither prospective fee-schedule rates nor estimates of the value of preventing disability. The eCRF should capture ICU days, hospital days, ventilation, discharge destination, rehabilitation, readmission where feasible and EQ-5D-5L so AIRCHILL-specific economic estimates can replace assumptions.

17. Go/no-go

ST-2 must show acceptable safety, no material EVT delay, reproducible thermal separation, coherent imaging/early neurological signal and a 90-day mRS distribution supporting a feasible pivotal trial.

← Clinical Investigation Package