AIRCHILL Helium Carrier-Gas Translational Study Protocol

Protocol · Version 1.0 · 23 August 2026

Helium as a carrier gas for faster respiratory brain cooling.

A staged translational protocol to test whether replacing a nitrogen-based carrier gas with helium materially improves AIRCHILL heat transfer while preserving ventilation, oxygen delivery and airway safety. The protocol is intentionally designed to answer the physical and physiological question first, before any human therapeutic-gas efficacy programme is considered.

Study synopsis

One variable changes: the carrier gas.

Primary hypothesis

Helium improves heat transfer

At matched FiO₂, airway pressure, minute ventilation, inlet-gas temperature and AIRCHILL operating mode, a helium-containing carrier gas will produce a faster fall in directly measured brain temperature than a nitrogen-based carrier gas.

Primary endpoint

Brain temperature change at 10 minutes

Difference between randomized groups in ΔTbrain from active cooling start to minute 10, measured at a prespecified intracerebral location and analysed blinded to allocation.

Key safety principle

Cooling cannot degrade ventilation

Any thermal advantage must occur without clinically relevant deterioration in oxygenation, CO₂ removal, airway pressures, pulmonary mechanics, haemodynamics or circuit performance.

Development use

Engineering selection study

The outcome selects the carrier-gas architecture for later AIRCHILL development. It is not a clinical efficacy study and cannot establish patient benefit.

Scientific rationale

Why helium deserves a controlled AIRCHILL experiment.

A 2025 porcine cardiac-arrest study reported higher brain-cooling efficiency when helium was incorporated into a respiratory cooling strategy, with approximately 0.8 °C brain-temperature reduction within ten minutes in the combined cooling condition. The result is directly relevant to AIRCHILL because helium changes gas density and thermal transport properties, but the published intervention was bundled and therefore does not isolate the carrier-gas effect. The next useful experiment is consequently a controlled head-to-head comparison in which the gas composition is the principal randomized variable.

Interpretation boundary. The working hypothesis is physical and physiological: helium may improve convective heat transfer through the respiratory cooling pathway. This protocol does not assume a pharmacological neuroprotective effect of helium and does not infer human benefit from the animal result.

Platform-integration rule. The helium study should not create bespoke monitoring, ECG, suction, defibrillation or connectivity subsystems merely for the experiment. Where suitable, use qualified calibrated laboratory equipment or the intended purchased/OEM modules and freeze their versions and interfaces in the protocol. The proprietary variable under study remains the AIRCHILL cooling/gas-path architecture; measurement-chain accuracy, gas-dependent sensor calibration and all interfaces that can influence the endpoint must still be verified.

Study design

Bench qualification first, then randomized porcine validation.

Stage	Purpose	Design	Proceed criterion
A · Bench	Verify that the AIRCHILL circuit can deliver helium mixtures safely and reproducibly.	Repeated-measures test-lung and thermal-lung model. Compare nitrogen-based carrier gas vs helium-containing carrier gas at matched FiO₂, minute ventilation, PEEP, tidal volume/pressure target and inlet temperature across the intended operating envelope.	No unacceptable sensor error, flow-control instability, pressure deviation, condensation problem, oxygen-delivery error or device alarm failure. Thermal advantage must be reproducible across repeated runs.
B · Acute preclinical	Isolate the carrier-gas contribution to brain cooling in a large-animal physiology model.	Randomized, controlled, parallel-group porcine study; 1:1 allocation. Both groups receive identical AIRCHILL cooling and ventilation except for carrier gas.	Primary endpoint and safety gate jointly satisfied.
C · Replication / arrest model	Confirm the effect in the intended resuscitation context if Stage B is positive.	Independent randomized cardiac-arrest/ROSC cohort using the selected gas architecture and prespecified replication criteria.	Directionally consistent thermal effect, acceptable safety and no material impairment of ROSC/haemodynamics.

Stage B randomized groups

Control

N₂-based AIRCHILL

Medical oxygen plus nitrogen-based carrier gas, titrated to the protocol FiO₂. AIRCHILL inlet temperature, humidification strategy, flow, ventilation target and cooling duration are fixed by protocol.

Experimental

He-based AIRCHILL

Medical oxygen plus helium-containing carrier gas, with the same target FiO₂, minute ventilation, airway-pressure strategy, inlet-gas temperature and cooling duration as control.

Randomisation: computer-generated 1:1 allocation in permuted blocks, concealed until the circuit is prepared. Blinding: operators cannot be fully blinded because gas handling differs, but temperature-endpoint adjudication, laboratory analysis and statistical analysis should be blinded to group. Sex: both sexes should be included where feasible and sex recorded as a prespecified covariate/exploratory interaction.

Endpoints

Thermal gain must be accompanied by physiological neutrality.

Tier	Endpoint	Definition
PRIMARY	ΔTbrain at 10 min	Brain temperature at minute 10 minus temperature immediately before active cooling, measured continuously and analysed from the prespecified intracerebral sensor.
KEY SECONDARY	Brain cooling rate	Slope in °C/min over 0–5, 0–10 and 0–20 minutes.
KEY SECONDARY	Time to thermal milestones	Time to −0.5 °C and −1.0 °C brain-temperature change where achieved.
SECONDARY	Core and regional temperature	Oesophageal/core, pulmonary/airway-adjacent and any validated regional surrogate to describe spatial thermal distribution.
SECONDARY	Heat-removal efficiency	Estimated heat extraction per unit time and, where technically valid, normalized to gas flow and temperature gradient.
SAFETY	Ventilation and gas exchange	PaO₂/FiO₂, PaCO₂, pH, SpO₂, end-tidal CO₂, tidal volume, minute ventilation, peak and plateau pressure, PEEP, dynamic/static compliance and airway resistance.
SAFETY	Haemodynamics	Heart rate/rhythm, MAP, arterial pressure, vasoactive support and clinically relevant instability.
DEVICE	Circuit performance	Delivered FiO₂, flow accuracy, pressure accuracy, temperature stability, condensation, obstruction, alarms, sensor drift and any gas-specific calibration error.

Statistical plan

Power the physical question, not a neurological outcome.

The initial large-animal comparison is planned around a conservative standardized thermal effect rather than the full effect reported in the 2025 bundled intervention. A working design of 24 evaluable animals, 12 per group, provides a practical confirmatory preclinical cohort for detecting a large between-group thermal effect while still allowing variance to be estimated for replication. The final animal number must be recalculated from bench/pilot variance and approved through the applicable ethical project evaluation before work begins.

Primary analysis: ANCOVA or linear model of 10-minute ΔTbrain with treatment group as fixed effect and baseline brain temperature as covariate.
Repeated thermal trajectory: mixed-effects model using all prespecified time points, with group × time interaction.
Safety: report effect estimates and confidence intervals, not only p-values. Prespecified physiological thresholds trigger individual stopping rules.
Missingness: reasons for missing temperature/safety data must be recorded; no undocumented deletion of failed runs or animals.
Multiplicity: one primary endpoint; secondary endpoints interpreted hierarchically/exploratorily unless a final statistical analysis plan specifies adjustment.
Analysis population: all randomized animals with protocol initiation; a per-protocol sensitivity analysis may be added for major technical deviations.

Sample-size governance. The 24-animal figure is a planning target, not an authorization. Stage A should estimate actual variance and device repeatability first. The animal application should then use that variance, the smallest scientifically meaningful ΔTbrain and the 3Rs principle to justify the minimum defensible number.

Safety, ethics and quality

The experiment only proceeds if ventilation remains safe.

Individual stopping

Physiological instability

Predetermine limits for severe hypoxaemia, uncontrolled hypercapnia/acidaemia, airway-pressure excursion, refractory hypotension/arrhythmia, circuit obstruction or other veterinarian-defined humane endpoints.

Study stopping

Device or gas-path signal

Pause enrolment after any unexpected serious device/gas-path event until root-cause review determines whether the study may safely continue.

3Rs

Bench before animal

Use test lungs and thermal models to eliminate unsuitable gas fractions and settings before any live-animal exposure. The in-vivo experiment should answer only questions that cannot be resolved on the bench.

Reporting

ARRIVE 2.0

Randomisation, blinding, exclusions, sample-size rationale, adverse events, protocol deviations and all prespecified outcomes should be documented in line with ARRIVE 2.0.

Regulatory boundary. Any live-animal work in the EU requires institutional and competent-authority approval under the applicable implementation of Directive 2010/63/EU before initiation. This public protocol is a scientific planning document, not an animal-use authorization.

Prespecified development decision

A positive study must change an engineering decision.

Advance helium architecture

Proceed if helium produces a clinically/engineering-relevant improvement in brain cooling, the confidence interval excludes a trivial effect, and ventilation/device safety remains within prespecified limits.

REFINE

Optimize fraction or flow regime

If thermal benefit is present but accompanied by sensor, pressure, gas-consumption or workflow penalties, return to bench optimization before replication.

NO-GO

Keep nitrogen-based carrier gas

If thermal benefit is small, inconsistent or offset by ventilation/device burden, do not add helium complexity to the AIRCHILL architecture.

Next protocol if GO. Replicate the selected helium configuration in a randomized cardiac-arrest/ROSC model with the same primary thermal measurements plus ROSC, haemodynamics and early neurological/biomarker outcomes. Only after reproducible preclinical performance and gas-path safety should a first-in-human feasibility protocol be drafted.

Evidence and governance sources

Protocol provenance.

The study is derived from the Medical Cooling evidence register and is designed to isolate the specific engineering question raised by the recent helium respiratory-cooling literature.

Helium cooling

Porcine respiratory-cooling study

PubMed PMID 41283153 — 2025 large-animal evidence that helium-containing respiratory cooling can increase brain-cooling efficiency.

Reporting

ARRIVE 2.0

ARRIVE Guidelines — design and reporting framework for rigorous in-vivo experiments.

EU animal research

Directive 2010/63/EU

EUR-Lex — EU framework for protection of animals used for scientific purposes.

Internal evidence map

Medical Cooling research register

Therapeutic-gas decision records — evidence grading and AIRCHILL transferability assessment.

Open full translational & human-performance pipeline →

Status: development protocol, Version 1.0. Before execution, the sponsor/research institution should freeze the device configuration, helium fraction, ventilation mode, temperature measurement hierarchy, humane endpoints, statistical analysis plan, animal-number calculation and authority-approved project application. Protocol amendments should be versioned rather than silently overwritten.