Paramedic Papers

Home Cardiac arrest EtCO2 trends predict ROSC cycle by cycle

Cardiac arrest

EtCO2 trends predict ROSC cycle by cycle

Cycle-to-Cycle Prediction of Return of Spontaneous Circulation Using End-Tidal Carbon Dioxide Dynamics in Out-of-Hospital Cardiac Arrest
Resuscitation · 2026-11 · Added 14 November 2026

A Taiwanese retrospective multicentre study of 1,480 patients with OHCA built a two-layer LSTM model on sequential EtCO2 that predicted ROSC well at event level (AUROC 0.876) but only 0.685 per two-minute cycle.

Study type
Modelling
Design
Multicentre retrospective study with LSTM prediction model, internally validated by 10-fold cross-validation
Participants
1,651 EtCO2 sequences from 1,480 adults with OHCA (583 ROSC events)
Main finding
Event-level AUROC 0.876 (Brier 0.140); cycle-level AUROC 0.685 across 17,029 two-minute prediction points, with systematic overestimation of ROSC probability (Brier 0.285)
Caveat
Retrospective, single-country dataset with internal validation only and poor cycle-level calibration
Bottom line
Hypothesis-generating

Why it matters in Australia (Medium relevance). EtCO2 monitoring is routine on Australian cardiac arrests, but this is a retrospective modelling study needing prospective, real-time validation before it guides cycle decisions on the road.

Learning goal

Weigh the value of EtCO2 trend monitoring for anticipating return of spontaneous circulation during resuscitation.

Open in dashboard Read the paper (DOI) PubMed

More

Related papers on Cardiac arrest