AIRCHILL Cardiac Arrest & Stroke Clinical Validation Plan
.acsv{–ink:#0B1226;–text:#3D4756;–muted:#6F7B8C;–line:#E4E8EE;–tint:#F6F8FC;–blue:#13A2EE;–mag:#C11C7A;font-family:Inter,-apple-system,”Segoe UI”,Arial,sans-serif;color:var(–ink)}.acsv *{box-sizing:border-box}.acsv .wrap{max-width:1180px;margin:auto}.acsv .hero{padding:118px 24px 78px;background:radial-gradient(900px 520px at 8% 0%,rgba(193,28,122,.45),transparent 60%),radial-gradient(900px 560px at 92% 100%,rgba(19,162,238,.38),transparent 62%),#080E20;color:#fff}.acsv .sec{padding:72px 24px}.acsv .light{background:#F6F8FC}.acsv .k{font-size:12px;font-weight:800;letter-spacing:.15em;text-transform:uppercase;color:#F3B8D8;margin:0 0 14px}.acsv h1{font-size:clamp(42px,6vw,72px);line-height:1.02;letter-spacing:-.04em;margin:0 0 22px;max-width:17ch}.acsv h2{font-size:clamp(30px,4vw,46px);line-height:1.08;letter-spacing:-.03em;margin:0 0 16px}.acsv h3{font-size:19px;margin:0 0 8px}.acsv p{line-height:1.68}.acsv .lead{font-size:18px;max-width:82ch;color:#C8D6E7}.acsv .sec .lead{color:var(–text)}.acsv .grid{display:grid;grid-template-columns:repeat(auto-fit,minmax(min(100%,260px),1fr));gap:16px;margin-top:28px}.acsv .card{background:#fff;border:1px solid var(–line);border-radius:18px;padding:23px;box-shadow:0 10px 28px rgba(11,18,38,.045)}.acsv .tag{font-size:11px;font-weight:800;letter-spacing:.11em;text-transform:uppercase;color:var(–mag);margin-bottom:10px}.acsv .table{overflow:auto;border:1px solid var(–line);border-radius:18px;margin-top:28px;background:#fff}.acsv table{border-collapse:collapse;width:100%;min-width:900px}.acsv th,.acsv td{text-align:left;vertical-align:top;padding:14px 15px;border-bottom:1px solid var(–line);font-size:13.5px;line-height:1.5}.acsv th{background:#0B1226;color:#fff;font-size:11px;text-transform:uppercase;letter-spacing:.055em}.acsv tr:last-child td{border-bottom:0}.acsv .call{margin-top:26px;border-left:4px solid var(–mag);background:#111A30;color:#C8D6E7;padding:20px 22px;border-radius:0 14px 14px 0;line-height:1.65}.acsv .call strong{color:#fff}.acsv .num{font-size:34px;letter-spacing:-.04em;font-weight:800;display:block;margin:4px 0 7px}.acsv .small{font-size:12.5px;color:var(–muted)}.acsv .dark{background:#0B1226;color:#fff}.acsv .dark p{color:#BFCBDA}.acsv a{color:#0B6FA6}.acsv .dark a{color:#8FD8FA}@media(max-width:760px){.acsv .sec{padding:54px 18px}.acsv .hero{padding-left:18px;padding-right:18px}}AIRCHILL clinical validation · version 1.0
Cardiac arrest & stroke: a programme designed to prove patient benefit.
The objective is not merely to show that AIRCHILL can lower temperature. The clinical programme is designed to show whether very early respiratory cooling, delivered without compromising ventilation, improves neurological outcome in carefully selected cardiac-arrest and ischaemic-stroke populations. The programme therefore separates safety and thermal performance from signal finding and from the final patient-outcome trial.
Executive answer
How many patients are likely to be needed?
Recommended planning range if Phase II supports an absolute improvement of roughly 7–9 percentage points in a prespecified favourable neurological outcome. A 5-point effect would require a much larger trial, around 4,400 patients under the conservative assumptions below.
Recommended planning range for a 90-day ordinal modified Rankin Scale trial if Phase II supports a common odds ratio around 1.35–1.40. If the true effect is closer to 1.30, a study of roughly 2,000 or more may be needed.
Randomised signal-finding work per indication is recommended before locking a Phase III effect size. This is the step that prevents the pivotal programme from being powered from an optimistic external signal.
Full protocol synopses
Two protocol-ready development documents.
Cardiac Arrest Clinical Trial Synopsis
Eligibility, randomisation, 90-day neurological endpoints, DSMB/interim framework, statistical analysis plan and recruitment scenarios through a ~2,300-patient base-case pivotal trial.
Ischaemic Stroke Clinical Trial Synopsis
Peri-EVT eligibility, no-delay workflow protection, blinded ordinal mRS, imaging core lab, DSMB/interim framework, statistical analysis plan and recruitment scenarios through a ~1,600-patient base-case pivotal trial.
Programme architecture
One platform, two indication-specific proof programmes.
| Stage | Cardiac arrest | Ischaemic stroke | Decision |
|---|---|---|---|
| 0 · Product safety / performance | Route-specific human airway safety and thermal-performance work after bench and large-animal qualification. Use a controlled intubated setting before emergency efficacy work. | Qualified operating envelope, ventilation stability, airway and cardiovascular safety. | |
| 1 · Emergency feasibility | ~80–120 selected OHCA patients, randomised or controlled, focusing on deployment, timing, ventilation continuity, thermal separation and safety. | ~60–100 EVT patients already requiring general anaesthesia / invasive ventilation; do not intubate solely for study participation. | Can the intended exposure actually be delivered in the target workflow? |
| 2 · Signal finding | ~300–500. Prespecified arrest phenotype, blinded 90-day neurological follow-up, biomarkers and complete treatment-clock reconstruction. | ~200–300. Randomise before or at EVT; imaging core lab, infarct-volume / penumbral-salvage endpoints plus blinded 90-day mRS. | Estimate the AIRCHILL-specific effect and select the pivotal population and thermal dose. |
| 3 · Confirmatory efficacy | Outcome-powered multicentre RCT, likely ~1,400–2,300 if the Phase-II effect is ≥7 pp; larger if smaller. | Outcome-powered multicentre RCT, likely ~1,300–1,600 for an ordinal mRS common OR around 1.35–1.40; potentially >2,000 for smaller effects. | Patient-relevant neurological benefit with prespecified estimand and blinded assessment. |
Cardiac arrest
Confirmatory OHCA study.
Preferred first efficacy population: adults with out-of-hospital cardiac arrest who remain unconscious and require invasive ventilation, with a deliberately narrow phenotype selected from Phase II. Initial shockable rhythm is a strong candidate because it creates a more homogeneous first pivotal population and mirrors the phenotype-enrichment logic used in PRINCESS2.
Start as early as operationally possible
AIRCHILL cooling integrated into invasive ventilation, initiated prehospital or immediately after airway control according to the final safety envelope. Record emergency call, EMS arrival, CPR, airway placement, randomisation, AIRCHILL start, first thermal effect, ROSC, hospital arrival and handover.
Guideline-based post-arrest care
Same ventilation and post-resuscitation standards without active AIRCHILL respiratory cooling. Fever prevention and downstream temperature management must be protocolised so the treatment contrast is interpretable.
Neurological survival
Blinded favourable neurological outcome at about 90 days, with the exact mRS threshold or an ordinal strategy frozen before the pivotal study. Mortality, full mRS distribution, CPC, EQ-5D-5L and cognitive / participation measures are key secondary outcomes.
Cooling cannot trade brain benefit for cardiovascular harm
Continuous ECG, haemodynamically significant arrhythmia, blood pressure, vasopressor dose, ventilation mechanics, gas exchange, pneumonia, bleeding, seizures and device-related events.
| Planning scenario | Control favourable outcome | AIRCHILL outcome | Absolute benefit | Approx. total randomised |
|---|---|---|---|---|
| Small but clinically important | 45% | 50% | +5 pp | ~4,400 |
| Moderate | 45% | 52% | +7 pp | ~2,250 |
| Large / PRINCESS2-scale | 45% | 54% | +9 pp | ~1,360 |
AIRCHILL planning calculation shown here: two-sided α=0.05, 90% power, 1:1 allocation, 5% primary-endpoint loss. For context, PRINCESS2 requires 1,022 participants to detect 45% versus 54% using 80% power, one-sided α=0.025 and 2.5% loss to follow-up. The external 9-point effect is a benchmark, not an AIRCHILL effect estimate.
Ischaemic stroke
Confirmatory LVO / EVT study.
Preferred first efficacy population: adults with acute anterior-circulation large-vessel-occlusion ischaemic stroke undergoing endovascular thrombectomy who already require general anaesthesia and invasive ventilation for clinical reasons. Randomisation should occur before or at the start of EVT so AIRCHILL tests early peri-reperfusion neuroprotection rather than late ICU cooling.
Peri-EVT AIRCHILL
Start once the airway is secured and continue through transport / EVT and a prespecified early post-reperfusion window. Capture time from last-known-well, imaging, randomisation, device start, reperfusion and first thermal effect.
Standard EVT care
Guideline-based thrombectomy, anaesthesia, ventilation and temperature management without active AIRCHILL respiratory cooling.
90-day ordinal mRS
The full modified Rankin Scale distribution is preferred over only mRS 0–2 because it uses information across the disability spectrum. Outcome assessors and the imaging core lab should be blinded.
Imaging can de-risk the pivotal study
Final infarct volume, infarct growth, penumbral salvage, reperfusion metrics and early NIHSS can establish a biological signal before the programme spends capital on a large functional-outcome trial.
| Ordinal mRS planning scenario | Common odds ratio for improvement | Approx. total randomised | Interpretation |
|---|---|---|---|
| Conservative | 1.30 | ~2,000+ | Small shift across the 90-day disability distribution. |
| Base planning case | 1.35 | ~1,600 | Reasonable pivotal planning anchor after a positive Phase II. |
| Stronger signal | 1.40 | ~1,300 | Still substantially more conservative than the large effects reported in small observational / pilot cooling studies. |
These ordinal figures are approximate planning values scaled from a contemporary thrombectomy-cooling protocol and should not be used as the final statistical calculation. The 2026 FOCUS protocol uses 258 patients for 80% power to detect a common OR of 1.9 on 90-day mRS. That is a very large assumed effect. A definitive AIRCHILL trial should use the actual Phase-II AIRCHILL mRS distribution and simulation-based proportional-odds / win-ratio sensitivity analyses before finalising N.
Rules that protect the proof
Design choices that matter as much as the patient count.
Time is part of the treatment dose.
Record every treatment clock and prespecify the maximum acceptable delay. A trial dominated by late treatment cannot answer whether very early cooling works.
Do not mix fundamentally different populations.
Shockable and non-shockable arrest, or different stroke territories / anaesthetic pathways, should be separated, stratified or prospectively enriched rather than rescued with post-hoc subgroup analysis.
Temperature alone is insufficient.
Store patient-side gas temperature, absolute humidity, flow, minute ventilation, FiO₂, pressure, duration, core / brain-temperature surrogates and rewarming so the actual respiratory thermal dose can be reconstructed.
Blind what can be blinded.
The device team cannot be blinded, but neurological outcome assessors, imaging core labs and preferably statisticians should be masked to assignment.
Use Phase II to prevent a mispowered Phase III.
A prespecified blinded sample-size re-estimation or adaptive design can be considered, but only with strict type-I-error control and regulator/statistician agreement.
Collect value data prospectively.
EQ-5D-5L, ICU / hospital days, ventilation days, rehabilitation, discharge destination, readmissions and long-term care should be collected alongside the confirmatory studies.
Evidence benchmarks
Sources used for this planning version.
PRINCESS2 design: PubMed 38417773 · TTM2: PubMed 34133859 · FOCUS protocol: PubMed 41984743 and NCT06485427 · COOLHEAD-2b: NCT07526649 · COTTIS: PubMed 41700739.
Status: development planning only. AIRCHILL has not yet established clinical efficacy in cardiac arrest or stroke. This page defines the studies intended to test and size that effect.
Operational package
From validation strategy to controlled study documents.
The programme now includes full development CIPs for cardiac arrest and stroke plus the shared Schedule of Assessments/CRF data model, DSMB Charter, Statistical Analysis Plan, screening/monitoring plan, safety-reporting plan, Investigator’s Brochure framework and EU/FDA submission dossier map.