Original Investigation 

Duration of Therapeutic Hypothermia After Out-of-Hospital Cardiac Arrest: The ICECAP Randomized Clinical Trial

William J. Meurer, Sharon D. Yeatts, Romergryko G. Geocadin, et al

JAMA Published Online: August 5, 2026

doi: 10.1001/jama.2026.10247

Key Points

Question  In comatose survivors of out-of-hospital cardiac arrest treated with therapeutic hypothermia at 33 °C, what duration of cooling maximizes neurological recovery?

Findings  In this multicenter, randomized, adaptive clinical trial including 1158 patients, the posterior probability that 6 hours was the shortest duration achieving maximal neurological recovery was approximately 0.5. The duration-response curve did not demonstrate improved efficacy with longer durations of therapeutic hypothermia.

Meaning  Among comatose survivors of out-of-hospital cardiac arrest treated with therapeutic hypothermia at 33 °C, increasing cooling duration did not improve neurological outcomes compared with shorter durations.

Abstract

Importance  Therapeutic hypothermia is widely used for neuroprotection following cardiac arrest, but clinical trials have not consistently demonstrated improved neurological outcomes, and the optimal duration of cooling remains uncertain.

Objective  To determine the duration of therapeutic hypothermia that maximizes neurological recovery in comatose survivors of out-of-hospital cardiac arrest.

Design, Setting, and Participants  Multicenter, randomized, adaptive-allocation clinical trial conducted at 71 hospitals in the US. Adults with out-of-hospital cardiac arrest who remained unconscious, achieved a target temperature less than 34 °C within 4 hours of cardiac arrest, and had a definitive temperature control device started were eligible. Patients were enrolled between June 2020 and June 2025.

Interventions  Therapeutic hypothermia at 33 °C with adaptive randomized allocation to cooling durations of 6, 12, 18, 24, 30, 36, 42, 48, 60, and 72 hours. The first 200 patients were randomized to 12-, 24-, and 48-hour durations in a 1:1:1 ratio. Subsequently, a response-adaptive randomization algorithm allocated preferentially to the groups most likely to be optimal and to best inform the duration-response curve separately within each rhythm type.

Main Outcomes and Measures  The primary outcome was neurological function at 90 days, measured using a weighted modified Rankin Scale score, analyzed using a bayesian duration-response model. The primary analysis estimated the posterior probability that each duration was optimal, wherein optimal indicates the shortest duration consistent with the best outcome observed at any duration.

Results  A total of 1158 patients were randomized (883 with nonshockable rhythms and 275 with shockable rhythms). Participants had a median age of 61 (IQR, 50-70) years and 39.6% were female. The trial met a prespecified stopping rule at the interim analysis. For the nonshockable rhythm cohort, the posterior probability that 6 hours was the shortest duration achieving the maximal mean weighted modified Rankin Scale score was 0.51. Results were similar in the shockable rhythm cohort. No differences were observed in secondary outcomes or mortality across cooling durations.

Conclusions and Relevance  Among comatose survivors of out-of-hospital cardiac arrest treated with therapeutic hypothermia at 33 °C, increasing cooling duration did not improve neurological outcomes.

Trial Registration  ClinicalTrials.gov Identifier: NCT04217551