COLD-DOSE Protocol · Human cold-air dose-response study

COLD-DOSE

Controlled human respiratory thermal-dose study

Protocol status: planning synopsis · version 0.1 · 24 August 2026
Evidence status: proposed mechanistic study. Exposure levels, cohort size and escalation criteria are ASSUMPTIONS requiring ethics, clinical, statistical and independent safety review.

Objective

Build a quantitative human dose-response model linking inspired temperature × absolute humidity × ventilation × duration to acute airway physiology, epithelial stress and recovery.

The purpose is not to discover the coldest air a person can tolerate. The purpose is to identify the respiratory exposure that maximises useful heat transfer while minimising airway stress and to establish measurable warning signals for later AIRCHILL studies.

Design

Randomised, controlled, within-participant crossover study with staged escalation and washout. Initial work should use healthy adults; an asthma/EIB susceptibility cohort can be added only after the healthy-participant safety envelope is characterised.

Planning sample

Initial healthy cohort: approximately 24–36 evaluable participants (ASSUMPTION). A smaller sentinel phase precedes full crossover exposure. Final N will be based on within-subject variability of the selected primary biomarker/physiology endpoint.

Exposure ladder

Candidate temperature levels:

1. thermoneutral/room-temperature control
2. 0 °C
3. −10 °C
4. −15 °C
5. −20 °C only after independent review of preceding levels

Temperature is never interpreted alone. Each exposure records absolute humidity, flow, minute ventilation, peak inspiratory flow and duration. A factorial or response-surface design should vary humidity and ventilation sufficiently to distinguish cooling from airway water loss.

Primary objective and endpoint

Primary objective: estimate the exposure-response curve for airway stress.

The final single primary endpoint should be selected after pilot reproducibility work. Candidate measures include change in FEV1/airway resistance or change in CC16 as a prespecified epithelial-stress biomarker. A composite should not be introduced merely to make the study positive.

Secondary endpoints

• serial FEV1, FVC and airway resistance
• small-airway function
• symptoms and bronchospasm
• SpO2 and, where justified, gas-exchange measures
• CC16 and prespecified epithelial/inflammatory biomarkers
• FeNO
• induced sputum in a mechanistic subset
• secretion/mucus characteristics where measurable
• nasal or other validated surrogate ciliary/mucociliary assessment
• heart rate, blood pressure and ECG surveillance appropriate to exposure intensity
• delivered gas temperature and absolute humidity at the participant interface
• calculated cumulative respiratory thermal/water-loss dose

Recovery is a formal endpoint

Measurements should not stop when cold exposure stops. Repeat airway physiology and selected biomarkers at prespecified recovery intervals, for example immediate, ~30–60 min, several hours and next-day where biologically appropriate.

A central development objective is to show and quantify reversibility: whether any small acute physiological or epithelial response returns toward baseline after rewarming/exposure cessation.

Randomisation and washout

Exposure order should be randomised within safety constraints. Extreme levels are gated rather than freely randomised before lower-dose review. Washout must be long enough for the selected physiological and biomarker endpoints to return to baseline. Baseline criteria should be prespecified before each subsequent exposure.

Safety escalation

An independent Safety Review Committee reviews sentinel data before opening −15 °C and again before −20 °C. Prespecified hold rules should include clinically important bronchospasm/FEV1 deterioration, oxygenation or ventilation abnormality, sustained cardiovascular signal, unexpectedly large epithelial biomarker response, serious symptoms or a pattern of delayed/non-reversible changes.

The exact numerical stopping thresholds must be finalised by investigators, ethics and the medical monitor; they are deliberately not invented from unrelated sports studies.

Statistical model

Use mixed-effects modelling with participant as a repeated/random effect and temperature, absolute humidity, ventilation and duration as exposure variables. Include interactions where biologically justified. The key output is a response surface with uncertainty intervals rather than a binary safe/unsafe temperature.

A candidate development metric is:

Thermal utility index = measured heat-removal/cooling performance ÷ prespecified airway-stress response.

This index is exploratory until validated and must not hide clinically important individual safety events.

Visit structure

Screening: health history, asthma/EIB history, baseline spirometry, contraindications.
Baseline session: repeatability and thermoneutral exposure.
Exposure sessions: randomised/gated temperature-humidity-ventilation conditions with continuous monitoring.
Recovery: serial physiology/biomarkers.
Final follow-up: confirm return to baseline and capture delayed events.

CRF minimum dataset

Participant ID; randomisation sequence; exposure temperature; interface temperature; absolute humidity; flow; minute ventilation; peak inspiratory flow; duration; cumulative dose; FEV1/FVC; airway resistance; small-airway metrics; FeNO; CC16; inflammatory markers; ciliary/mucociliary metric; symptoms; SpO2; ECG/vitals; recovery values; adverse events; stopping criteria; deviations.

Evidence boundary

Existing controlled cold-air studies demonstrate that humans have undergone short cold-air challenges into approximately the −20 °C range and, in small historical physiology experiments, colder brief exposures. They also show that high ventilation and dry air can provoke bronchoconstriction or epithelial/inflammatory responses. These studies justify a staged mechanistic programme; they do not establish −20 °C or any other value as a universal human safety threshold.

Relevant anchors include the 2026 systematic review of cold-air challenge, Eklund et al. (PMID 36053365), Kennedy et al. (PMID 29178677), Frisch et al. (PMID 35391634), Clary-Meinesz et al. (PMID 1305479), and the completed CLARINET registry NCT07220928.

Translation to AIRCHILL

COLD-DOSE identifies sensitive endpoints and dose relationships. It does not replace the AIRCHILL prolonged endotracheal animal programme or the subsequent product-specific intubated human study because route, duration, conditioning and clinical population differ.

Registration-ready outcomes and time points

Proposed primary outcome (ASSUMPTION): within-participant change in the selected airway-stress measure across the prespecified respiratory thermal-dose conditions. The pilot must freeze either a physiological endpoint (for example maximum post-exposure FEV1 change / airway resistance) or an epithelial endpoint (for example CC16 change) before the confirmatory crossover portion; both must not be retrospectively promoted to co-primary outcomes.

Key secondary outcomes: serial FEV1/FVC and airway resistance; small-airway function; symptoms/bronchospasm; SpO2; FeNO; CC16 and prespecified epithelial/inflammatory markers; mucociliary/ciliary surrogate; cardiovascular measures; and recovery to the participant’s baseline. Exposure variables are patient-interface temperature, absolute humidity, minute ventilation, peak inspiratory flow and duration.

Time points: pre-exposure baseline; continuous monitoring during exposure; immediate post exposure; approximately 30–60 min; later same-day sampling where biomarker kinetics justify it; and next-day/final follow-up for recovery and delayed events. Exact specimen times will be locked after biomarker-method review.

Schedule of assessments

Sample-size plan

The current 24–36 evaluable participants is an ASSUMPTION appropriate for protocol development, not a final powered claim. A sentinel/pilot subset will estimate within-person standard deviation, carry-over and biomarker reproducibility. The confirmatory crossover N will then be calculated from the minimum clinically/biologically relevant within-participant change in the frozen primary endpoint, its observed within-person variance, the final number of conditions, multiplicity strategy and anticipated dropout/non-evaluable sessions. The calculation and simulation code should be archived before registration of the confirmatory portion.

Safety Review Committee charter · minimum content

• independent chair plus pulmonary/critical-care expertise and statistical support;
• review after sentinel thermoneutral/0/−10 °C exposure and before opening −15 °C;
• second formal review before −20 °C;
• review all SAEs, exposure-related AEs, individual pulmonary trajectories, oxygenation, cardiovascular signals, CC16/epithelial responses, delayed recovery and protocol/device deficiencies;
• authority to continue, modify, pause, de-escalate or terminate an exposure level;
• written decision and rationale retained in the trial master file.

Hold criteria categories: clinically meaningful bronchospasm or lung-function deterioration, oxygenation/ventilation abnormality, sustained cardiovascular abnormality, severe symptoms, unexpected biomarker pattern, delayed/non-reversible change, or any SAE plausibly related to exposure. Numerical thresholds remain DATA GAP until the final population, equipment and medical-monitor plan are fixed.

Decision gate

A candidate AIRCHILL clinical exposure is selected only when the response surface, recovery data, preclinical endotracheal evidence and engineering performance jointly support it. The current −10 → −15 → −20 °C AIRCHILL escalation remains a provisional development assumption, not an output predetermined by this study.