Resuscitation From Out-of-Hospital Cardiac Arrest: Reliable Association of EtCO2 with ROSC

Introduction

Out-of-hospital cardiac arrest (OHCA) represents a critical emergency with high mortality rates worldwide. Early and accurate assessment of resuscitation effectiveness is essential for improving patient outcomes. One emerging tool in this setting is continuous monitoring of exhaled end-tidal carbon dioxide (EtCO2), which reflects pulmonary blood flow and ventilation status during cardiopulmonary resuscitation (CPR). EtCO2 trajectories may indicate the likelihood of return of spontaneous circulation (ROSC), guiding clinical decisions at the critical bedside.

Despite known associations between EtCO2 levels and resuscitation outcomes, the exact duration of EtCO2 monitoring required to reliably differentiate patients who will regain ROSC from those who will not remains uncertain. Understanding this timing could optimize clinical protocols and resource utilization during emergency resuscitation.

Methods

This investigation conducts a secondary analysis of data from the Pragmatic Airway Resuscitation Trial (PART), a large cluster-randomized study comparing endotracheal intubation and laryngeal tube airway management strategies in OHCA resuscitation scenarios. The analysis focuses on the EtCO2 data collected during resuscitation efforts.

The study categorized patients according to two key criteria: whether the cardiac arrest was witnessed versus unwitnessed, and initial EtCO2 values categorized as low (≤30 mm Hg), moderate (31–49 mm Hg), or high (≥50 mm Hg). Group-based trajectory modeling (GBTM) identified distinct latent patterns of EtCO2 changes over time, grouping patients into upward or downward EtCO2 trajectory classes.

To control for baseline differences such as patient age, sex, race, initial heart rhythm, location of arrest, and presence of bystander CPR, inverse probability of treatment weighting was applied. Using weighted pooled logistic regression, risk ratios (RRs) were computed to compare the probability of ROSC between upward and downward EtCO2 trajectory groups. The analysis also pinpointed the earliest minute during resuscitation when confidence intervals (CIs) for these trajectories no longer overlapped, indicating a reliable distinction between outcomes.

Results

Out of 1168 patients with available EtCO2 data, 452 (38.6%) were witnessed arrests and 716 (61.1%) were unwitnessed. The study population was predominantly male (63.5%), median age 65 years, majority White race (51.3%), with most arrests occurring in nonpublic settings (85.4%). The overall ROSC rate was 18.2%, with a higher rate in witnessed arrests (30.5%) compared to unwitnessed (10.5%).

For witnessed arrests:

• Patients with low initial EtCO2 (≤30 mm Hg) showed a significant separation of ROSC probability between upward and downward trajectories at 8 minutes (RR 3.06; 95% CI 1.49–6.71).
• Those with moderate initial EtCO2 (31–49 mm Hg) showed separation at 12 minutes (RR 1.95; 95% CI 1.23–3.48).
• Patients with high initial EtCO2 (≥50 mm Hg) showed this distinction later at 21 minutes (RR 2.12; 95% CI 1.30–3.73).

For unwitnessed arrests, the difference in trajectories became reliably distinguishable earlier, at 7 minutes (RR 3.56; 95% CI 1.53–10.37), regardless of initial EtCO2 grouping.

Discussion

This study demonstrates that the time needed to confidently predict ROSC based on EtCO2 monitoring varies according to whether the cardiac arrest was witnessed and the patient’s initial EtCO2 values. Generally, 7 to 21 minutes of continuous EtCO2 monitoring during resuscitation is necessary to differentiate patients on likely recovery trajectories.

These findings support dynamic monitoring of EtCO2 trajectories rather than reliance on static EtCO2 values alone. The upward or downward pattern offers prognostic insights that can inform resuscitation strategies, such as continuation, modification, or termination of efforts.

Clinicians should consider that lower initial EtCO2 requires a shorter observation time to predict ROSC compared to higher initial EtCO2. Moreover, witnessed arrests—which are often treated earlier and may have more reversible causes—require slightly longer monitoring to differentiate outcomes.

Limitations include lack of information on other adjunctive resuscitation modalities and generalizability beyond the study settings. Further research should explore integration with other clinical parameters and technologies, including advanced hemodynamic monitoring and machine learning prediction models.

Conclusion

End-tidal CO2 trajectory monitoring during out-of-hospital cardiac arrest resuscitation provides valuable early prognostic information. Depending on initial conditions, between 7 and 21 minutes of EtCO2 data are required to reliably distinguish patients with return of spontaneous circulation from those without. Tailoring resuscitation approaches based on these dynamic EtCO2 patterns may improve patient outcomes in this critical setting.

Clinical Implications

Emergency medical providers and critical care personnel should incorporate continuous EtCO2 monitoring into standard OHCA resuscitation protocols. Interpretation of upward versus downward EtCO2 trends, with attention to initial EtCO2 and witness status, can facilitate timely clinical decision-making. Education on these principles can enhance emergency response effectiveness and patient survival chances.