Capnography Missed PaCO2 in Most Prehospital Intubations

Key Takeaways

• In a single Southern Finland HEMS cohort of critically ill patients after prehospital emergency anesthesia and intubation, most had a PaCO2-EtCO2 gradient above the assumed threshold.
• Overall agreement between arterial and end-tidal carbon dioxide was wide rather than fixed around the guideline-assumed gradient.
• Discordance remained common across every diagnostic group studied, with the highest pattern among patients intubated after ROSC.
• The authors concluded that capnography alone was not accurate enough to guide normoventilation in this prehospital setting.

Prehospital ventilation targets for intubated critically ill patients often rely on end-tidal carbon dioxide capnography (EtCO2) as a practical surrogate for arterial carbon dioxide partial pressure (PaCO2), yet shock, ventilation-perfusion mismatch, and mixed underlying diagnoses can disrupt that relationship during early resuscitation. In a retrospective study of prehospital PaCO2-EtCO2 agreement, investigators reviewed a retrospective observational registry from 1 January 2014 through 31 December 2021 at FinnHEMS 10, a physician-staffed helicopter emergency medical services (HEMS) unit in Southern Finland, focusing on critically ill patients who underwent prehospital emergency anesthesia (PHEA) and intubation.

Of 2203 anesthetized and intubated patients who were eligible, 670 (30.4%) had a post-intubation arterial blood gas (ABG) sample before hospital arrival and entered the analysis. Patients intubated during ongoing out-of-hospital cardiac arrest (OHCA) were excluded, while patients intubated after return of spontaneous circulation (ROSC) were included; diagnostic groups were OHCA after ROSC, trauma, intoxication, neurologic conditions, and other. ABG testing was performed at physician discretion with a point-of-care device, each ABG was paired with the nearest EtCO2 value within 5 minutes, and only the first arterial sample was used when more than one was available. Bland-Altman analysis defined agreement by the bias and limits of agreement between paired PaCO2 and EtCO2 values.

The comparison showed an overall PaCO2-EtCO2 bias of 1.7 kPa (SD 1.6). Overall limits of agreement were reported as -1.4 to 4.9 kPa. Agreement remained broad across all diagnostic groups.

The PaCO2-EtCO2 gradient exceeded the guideline-assumed threshold in 530/670 patients (79.1% [75.9-82.0]). The highest subgroup endpoint was 84.2% among patients assessed after ROSC, while the lowest reported endpoint was 74.1% in neurologic conditions; trauma, intoxication, and other diagnoses also remained above the threshold. The authors also noted that bias was higher in the OHCA and other groups and appeared to rise with overall carbon dioxide levels.

The authors described several constraints on interpretation, including the retrospective single-unit design, ABG sampling at physician discretion with only a subset of intubated patients entering the analysis, and the lack of simultaneous PaCO2 and EtCO2 measurement. They also noted that time from intubation to ABG sampling was unavailable, diagnostic categories were heterogeneous, and monitor-defibrillator equipment changed over the multiyear study period.

Clinician Questions

Does the prehospital EtCO2-PaCO2 mismatch analysis apply to patients intubated during ongoing OHCA?

No. The analysis excluded patients intubated during ongoing out-of-hospital cardiac arrest, while patients intubated after return of spontaneous circulation were included as the OHCA subgroup, so the reported mismatch findings apply to post-ROSC patients rather than active arrest intubation.

How were arterial and end-tidal CO2 values matched in the Southern Finland PHEA cohort?

Investigators used arterial blood gas samples obtained after intubation and before hospital arrival, paired each sample with the closest end-tidal carbon dioxide reading in the prehospital record, and used only the first arterial sample when more than one prehospital sample was available.

What mechanisms did the authors discuss for a wider PaCO2-EtCO2 gap during prehospital anesthesia?

The authors discussed ventilation-perfusion mismatch, hypotension reducing pulmonary circulation, and inadequate ventilation or atelectasis as contributors to a wider PaCO2-EtCO2 gap during prehospital anesthesia. They also observed higher bias in the OHCA and other groups, while noting that the reasons for that pattern remained uncertain.