EBM Review Series: End-Tidal CO2 Measured at Triage Outperforms Standard Vital Signs – Potential Utility and Value of ETC02
VACEP Evidence-Based Medicine for General Emergency Physicians Series
Authors: Samuel Reed, MD and Moira Smith, MD, MPH | UVA Medical Center
Reviewers: Bianca Mayfield, DO PGY-3 & Donald Engle, MD | Naval Medical Center Portsmouth
The VACEP Evidence-Based Medicine Review Series allows Virginia emergency medicine residents and attendings to share and analyze a recent peer-reviewed clinical study. You can also read the full article, “End-tidal carbon dioxide measured at emergency department triage outperforms standard triage vital signs in predicting in-hospital mortality and intensive care unit admission”, in Volume 30, Issue 8 of Academic Emergency Medicine.
THE CASE
A patient is brought to the emergency department (ED) by emergency medical services (EMS) with concern for sepsis. The patient has a history of urosepsis but is oriented with vital signs in the normal range. As a result, there was not a sepsis alert prior to this patient’s arrival. Initially, the patient appears mildly fatigued with vital signs significant for a heart rate of 96 beats per minute and a blood pressure of 106/62 mmHg. You - the physician - are deciding if it is appropriate to sepsis alert the patient now, or if you should first collect more data.
What other information could be used to better inform this decision?
BACKGROUND
Carbon dioxide (CO2) is a metabolic byproduct, and its clearance depends on ventilation and circulation. Exhaled CO2 is reduced in hypotensive states, including decreased cardiac output or even dehydration. It is also reduced during tachypnea, which could reflect a range of pathology from metabolic acidosis to pain or anxiety.
Today in the ED, end tidal CO2 (ETCO2) is utilized as a measure of ventilation during procedural sedations and during other periods in which the airway is potentially tenuous. It is used during resuscitation to grossly indicate the return of spontaneous circulation. Increasingly, data has shown the additional utility of ETCO2 as a diagnostic tool in predicting severe illness.
In 2019, Weiss et al. conducted a retrospective study which included all patients who received a pre-hospital sepsis alert by their local EMS agency and were transferred to their Level 1 trauma center. They utilized prehospital ETCO2 measurements collected by EMS in addition to laboratory values and other data collected post-arrival. They found an inverse linear relationship between lactate and ETCO2 with an R value of -0.45. At a cutoff of ETCO2 < 25, a “positive” measurement predicted lactate of >4 with 63% sensitivity and 80% specificity, with an AUC of 0.73, indicating some utility of ETCO2 in predicting acidosis.
Another study by Willis et al. in 2022 found that decreased ETCO2 was even more predictive of mortality than elevated lactate, decreased systolic blood pressure (SBP), and other more conventional evidence of hemodynamic instability in the trauma population. This was a retrospective study performed at a Level 1 trauma center. Their study population was all patients for whom a trauma alert was called and who had an available ETCO2 level. They constructed receiver operator characteristic (ROC) curves to assess the relationships of multiple independent variables with mortality including ETCO2, lactate, SBP, and acidosis. They found that the area under the curve (AUC) for ETCO2 in predicting mortality was 0.75, which was significantly more predictive than lactate with an AUC of 0.66, SBP (0.58), and pH (0.56).
Clearly, there is evidence to suggest a potential utility for ETCO2 in predicting critical illness, but how exactly can a measurement of ETCO2 be applied to our hypothetical patient to guide initial care? A study recently conducted by Laddle et al. may provide answers to these questions by further comparing ETCO2’s predictiveness of multiple outcomes against the predictiveness of every standard vital sign, using prospectively collected data to which the treatment team was blinded.
STUDY SUMMARY
Laddle et al. conducted a prospective observational study performed at a Level 1 trauma center in Florida which included 1091 patients who presented either by EMS or through triage. To meet the inclusion criteria patients needed to be both 18 or older and to have been allocated an emergency severity index (ESI) score of II, III, or IV. These patients were seen by a member of the investigator team either in triage or, if brought directly into the department by ambulance, in their ED rooms, and ETCO2 was measured by a standardized process concurrently with other vital signs. If an ETCO2 had previously been recorded by a nurse or by EMS, then that value was used. The treatment team was blinded to ETCO2 measurements.
OUTCOME MEASURES:
The primary outcome of their study was in-hospital mortality, and more accurately, ETCO2’s predictiveness of in-hospital mortality based on an ROC curve which was constructed.
An ROC curve was additionally constructed to assess the relationship between initial ETCO2 and ICU admission, as a secondary outcome.
Furthermore, ROC curves were constructed to assess more traditional “vital signs” - including systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), respiratory rate (RR), oxygen saturation (SpO2), and temperature - in their ability to predict the primary and secondary outcome.
Backward logistic regression was performed to better isolate the relationship between ETCO2 and in-hospital mortality, adjusting for other vital signs, and the results of this were expressed as an odds ratio (OR).
As a tertiary outcome, ETCO2 measurements were correlated with other measurements of metabolic acidosis using a rho statistic.
RESULTS
Compared to standard vital signs, they found ETCO2 to be the most predictive of in-hospital mortality with an AUC of 0.82 (0.72 - 0.91). The AUC for all standard vital signs was lower, although in four cases there was overlap of the confidence intervals. SBP had an AUC of 0.77 (0.67 - 0.86), DBP had an AUC of 0.7 (0.59 - 0.81), HR had an AUC of 0.76 (0.66 - 0.85), and RR had an AUC of 0.59 (0.46 - 0.73). After backward logistic regression, ETCO2 was still found to be the most predictive of in-hospital mortality with an OR of 0.88 (0.83 - 0.93, p < 0.001), and aside from ETCO2 only pulse was significantly predictive of in-hospital mortality after backward logistic regression analysis with an OR of 1.023 (1.005 - 1.041, p = 0.013). Similarly, ETCO2 was found to be most predictive of the secondary outcome of ICU admission with an AUC of 0.75 (0.67 - 0.80). Furthermore, ETCO2 was found to significantly correlate with serum lactate (rho -0.26, p < 0.001), bicarbonate (rho 0.33, p < 0.001) and anion gap (rho -0.2, p < 0.001).
Laddle et al. attempted to characterize the test characteristics of ETCO2 for predicting the primary outcome at multiple cutoff points, and notably found that at a cutoff of 28 mmHg, low ETCO2 was 77% sensitive and 79% specific for in-hospital mortality with a negative predictive value of 99% and a positive predictive value of 8%.
About the EBM Review Series
This is a literature review series started by the University of Virginia’s Josh Easter, MD, MSc, a VACEP board member working to connect the academic community in Virginia. We invite each residency in Virginia (and D.C.) to create a faculty/resident team to submit and review articles. Sign up to submit one.
Goals
Provide a brief monthly synopsis of a high yield article germane to the practice of emergency medicine for distribution to all VACEP members
Provide an opportunity for a peer reviewed publication and invited presentation for faculty and trainees
Foster an academic community focused on evidenced based medicine for emergency medicine residency programs in the region
STRENGTHS, LIMITATIONS, & IMPLICATIONS
This study was carried out in a single ED and the staff collecting ETCO2 measurements had received a 1-hour training session prior to data collection. As a result, there are potential limitations to the generalizability of their results, and achieving their test characteristics may require a resource investment in further staff training.
There is not one specific process that can be tied causally to mortality or ICU admission associated with lower ETCO2 levels.
While this study indicates potential utility of ETCO2 in predicting illness severity, it does not provide insight into any improvement in care that may be gleaned by utilizing this data.
CONCLUSIONS
The authors of Laddle et al. conclude that ETCO2 could be used to accurately predict their primary and secondary outcomes of in-hospital mortality and ICU admission respectively, and that it was more predictive of these outcomes than standard vital signs. They further comment on its potential utility as a triage vital sign, indicating that it could impact early medical decision making. In recent years, other authors have conducted similar studies and drawn similar conclusions. For instance, in their 2023 study, Tortum et al. suggest that ETCO2 could be used a triage vital to determine the triage category of patients.
END OF CASE
In addition to standard vital signs, you use capnography to measure your patient’s ETCO2 and find it to be 27 mmHg, indicating significant risk of mortality and need for resuscitation. You call a sepsis alert, and quickly afterward the patient has multiple points of IV access. Fluids and antibiotics are started, and a subsequent venous blood gas reveals a pH of 7.28 and a lactate of 3.7. Your patient ultimately remains stable and is admitted to the hospital with a diagnosis of pyelonephritis.
SOURCES
Weiss, Steven J., et al. "Sepsis alerts in EMS and the results of pre-hospital ETCO2." The American Journal of Emergency Medicine 37.8 (2019): 1505-1509.
Willis, Robert G., et al. "Prehospital end-tidal CO2: a superior marker for mortality risk in the acutely injured patient." The American Surgeon 88.8 (2022): 2011-2016
Ladde, Jay G., et al. "End‐tidal carbon dioxide measured at emergency department triage outperforms standard triage vital signs in predicting in‐hospital mortality and intensive care unit admission." Academic Emergency Medicine 30.8 (2023): 832-841.
Tortum, Fatma, et al. "A New Vital Sign in Determining the Triage Category in Emergency Department Presentations: End-Tidal Carbon Dioxide." European Journal of Therapeutics29.4 (2023): 689-697.
