Controlled cooling, built into ventilation.

Germany sees an estimated 54,000–67,000 EMS resuscitation attempts per year, and roughly 16,000–20,000 patients are admitted to hospital with return of spontaneous circulation. For the arithmetic below we use an illustrative eligible cohort of 7,500 patients per year. This 7,500 figure is a modelling input for a potentially narrower early-cooling population, not a registry-derived estimate of current eligibility.

A 2024 observational registry analysis of 33,933 comatose adult cardiac-arrest patients found mild therapeutic hypothermia to be independently associated with CPC 1–2 survival with an adjusted OR of 1.60 (95% CI 1.49–1.72). Translating that adjusted odds ratio onto the observed 14.0% CPC 1–2 rate in the non-MTH group gives an illustrative rate of about 20.7%, an absolute difference of roughly +6.7 percentage points. Applied to the 7,500-patient modelling cohort, that would equal about 500 additional favourable neurological outcomes per year. At the existing €25,000–€75,000 value sensitivity, that corresponds to roughly €12.5–€37.5m gross per annual cohort.

Important: this is a signal translation, not a causal treatment estimate. The study was retrospective and observational, included both OHCA and IHCA, evaluated MTH after ROSC rather than prehospital AIRCHILL, and the treatment groups differed in many baseline and care variables. The crude CPC 1–2 rates were 33.3% versus 14.0% (+19.3 percentage points), but we deliberately do not extrapolate that crude difference as the model effect.

The COTTIS matched-pair analysis reported favourable 90-day outcome (mRS 0–2) in 68.2% (15/22) of hypothermia patients versus 29.5% (13/44) of matched standard-care patients: an observed absolute difference of +38.7 percentage points, with unadjusted OR 5.1 (95% CI 1.69–15.38). If that signal were translated purely mathematically to the 18,809 annual German thrombectomy system cohort used above, it would equal about 7,279 additional patients with mRS 0–2 per year. At the €15,000–€35,000 health-economic sensitivity used in this model, the gross arithmetic would be about €109–€255m per annual cohort before intervention and downstream costs.

This is an upper-bound signal scenario, not a forecast. COTTIS was a small single-centre, non-randomized matched-pair analysis of 66 patients; the authors explicitly describe the effect size as surprisingly high and requiring critical consideration and randomized validation. The study used rapid transnasal cooling, not AIRCHILL, and a real eligible AIRCHILL population would be narrower than all German thrombectomy patients.

For stroke, the scientifically stronger system endpoint is a patient-relevant shift in disability — commonly 90-day mRS or mRS 0–2 — rather than a separate speculative “lives saved” calculation. AIRCHILL has treated 0 patients; no clinical efficacy or patient-safety data exist for AIRCHILL.

Germany recorded 654,241 live births in 2025. Epidemiological literature commonly places HIE in high-income settings at about 1.5 per 1,000 live births, corresponding to an upper system frame of roughly 980 infants per year. The actual population eligible for therapeutic hypothermia is smaller because gestational age, encephalopathy severity, timing, diagnostic criteria and contraindications must be applied.

There is no single German lifetime tariff for HIE-related disability. We therefore use two European anchors. A register-based Danish cerebral-palsy study estimated discounted attributable lifetime costs of about €0.80–€0.86m per person in historical euros. A 2026 population-based Spanish CP study, expressed in 2023 euros, reported average childhood societal costs of about €102,000 per child per year; severe motor impairment was associated with almost twice the annual cost. For sensitivity analysis we therefore model €0.8–€2.0m lifetime societal burden per avoided severe lifelong neurodevelopmental disability. This is not a German reimbursement claim.

To avoid treating mortality and disability as the same economic endpoint, we additionally assume that only 50–100% of the incremental favourable outcomes represent a shift from moderate/severe lifelong disability to a substantially better functional trajectory.

AIRCHILL in its current development form is not validated for neonates and this is not a current product claim. Neonatal use would require independent engineering for neonatal tidal volumes and gas flows, temperature control, humidification, airway safety, monitoring, dosing logic, transport integration, paediatric risk management, regulatory strategy and prospective clinical evaluation.

A prospective Japanese multicentre registry analysis found in-hospital mortality of 21.5% with active cooling plus rehydration versus 35.5% with rehydration alone among 231 severe heat-illness patients — an observed absolute difference of 14 percentage points. After multivariable adjustment, rehydration-only care was associated with higher odds of death than active cooling (aOR 3.29; 95% CI 1.21–8.90). Translated only as an upper signal scenario, that is approximately 14 additional survivors per 100 comparable severe patients.

Not a causal product estimate: this was observational, cooling methods were heterogeneous, and AIRCHILL was not studied. The 14-point difference is therefore a best-case signal boundary, not a forecast.

For exertional heat stroke, immediate whole-body water immersion remains the best-supported rapid cooling approach in the ILCOR review and should not be displaced by a speculative respiratory device. A future respiratory-cooling system could only be studied as an adjunct or alternative in selected severe/intubated cases, during transport, or when established whole-body cooling is not practical. Classic heat stroke — often affecting older, multimorbid patients — is the area in which the optimal cooling technique remains especially uncertain.

AIRCHILL has treated 0 heat-stroke patients. It has no clinical efficacy or patient-safety evidence for heat stroke and no current product claim for this indication. Any heat-stroke programme would require dedicated bench testing, thermal-control limits, stopping rules, airway and pulmonary safety work, clinical protocol development and prospective evaluation.

German registry-based epidemiology supports an order of magnitude of about 11,400 moderate-to-severe TBI cases per year. For the arithmetic below we deliberately use a narrower 7,500-patient operational cohort to reflect exclusions for injury pattern, haemodynamic instability, bleeding risk, lack of invasive ventilation, contraindications and practical reachability. The 7,500 figure is a modelling input, not a published eligibility estimate.

In the prespecified LTH-1 subgroup with initial intracranial pressure ≥30 mmHg, favourable outcome occurred in 60.82% with prolonged mild hypothermia versus 42.71% in controls, an observed absolute difference of +18.11 percentage points. Purely mathematically applied to the 7,500-patient modelling cohort, that would equal about 1,358 additional favourable outcomes. The first-year cost sensitivity would be approximately €27–€68m; a long-term model would be roughly €340m–€1.02bn per annual cohort.

Upper-bound signal, not forecast: the overall LTH-1 trial did not show a statistically significant primary outcome benefit. POLAR found no benefit from early prophylactic hypothermia, while Eurotherm3235 showed worse functional outcome despite lowering intracranial pressure. Cochrane judged the overall TBI hypothermia evidence highly uncertain. These conflicting results argue for careful phenotype selection, bleeding and haemodynamic safeguards, controlled rewarming and patient-relevant endpoints — not for a broad treatment claim.

About 44,000 sports-related concussions per year are reported as an order-of-magnitude German figure. We use this sports population because it most closely resembles the participants studied in the existing selective-cooling literature, rather than extrapolating from the much larger and more heterogeneous total mild-TBI population.

In a non-randomized Swedish elite-ice-hockey cohort, median return to play was 9 versus 13 days after early head-and-neck cooling, and prolonged absence beyond 14 days occurred in 24.7% versus 43.7%. The observed 19-percentage-point difference would mathematically correspond to about 8,360 favourable trajectories if reproduced across 44,000 comparable cases. At the €3k–€15k sensitivity, that is roughly €25–€125m gross in the first year.

Not a forecast: the cohort was not randomized. A newer multicentre randomized study did show greater reduction in SCAT5 symptom severity with selective head-and-neck cooling, but also reported more adverse events in the cooling arm. These studies address symptoms and return to activity, not mortality or prevention of severe neurological disability.

Concussion is not a current AIRCHILL use case. The positive evidence concerns external selective cooling systems, not cooled ventilation gas. Concussion may justify a separate future non-invasive neuroprotection research pathway, but it should not be presented as an AIRCHILL efficacy or indication claim.