EtCO2 vs. Standard Triage Vitals in Predicting In-Hospital Mortality and ICU Admission

/

Ladde JG, Miller S, Chin K, 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. Acad Emerg Med 2023; 10.1111/acem.14703

  • P: 1094 patients, triaged as ESI 2-4, presenting to a single center, Level 1 trauma center in Florida by either lobby or EMS

  • I: EtCO2 Measurement

  • C: Traditional vital signs (HR, BP, RR, SpO2)

  • O: EtCO2 outperformed traditional vital signs in predicting in hospital mortality, ICU admissions and had a significant correlation to anion gap, serum lactate and HCO3

Background

Boarding of admitted patients in the ED and subsequent overcrowding of ED’s continues to plague hospitals in the United States and Internationally.  The Covid-19 pandemic exacerbated an already growing problem regarding capacity management and patient flow. In this current climate, the Emergency Physician’s responsibilities continue to shift toward the front-end of the process, mainly patients waiting to be seen in the lobby.  As such, identifying sick patients in a timely manner and utilizing additional resources to predict patients at risk of clinical deterioration will be paramount moving forward.

The current triage model in the Emergency Department utilizes the ESI algorithm that considers the need for immediate lifesaving interventions, high risk situations, vital signs and resources. Traditional vital signs, (HR, BP, RR, SpO2), quickly provide non-invasive measures of a patient’s condition and are critical to risk stratification of patients arriving in the ED.  However, there is some suggestion in the literature that the reproducibility of vital signs among healthcare providers is inconsistent, with studies demonstrating that vital signs do not reliably capture the severity of illness in certain patient populations. (1) Identifying those critically ill patients who might otherwise go undetected using traditional vital signs is a challenge in Emergency Medicine.

EtCO2 is a valuable tool in many medical specialties.  Utilized in the pre-hospital setting, ICU and the Operating Room, it is a non-invasive measurement that reflects a patient’s current state of metabolism, circulation and ventilation. (5,6)   It is commonly used to confirm endotracheal tube placement after intubation or to monitor respiratory effort during procedural sedation. It’s association with lactate and anion gap make it an attractive option for identification of critically ill patients. It’s reflection of underlying metabolism, circulation and ventilation may serve well to identify patients with early dysfunction of physiologic functions. The object of this prospective study was to assess the ability of EtCO2 to predict in-hospital mortality and ICU admission compared to standard vital signs at ED triage in an undifferentiated ED population.

Methods

This was a prospective observational study enrolling all patients who presented to the ED through the lobby and via EMS. This was a convenience sample, as patients were only enrolled when investigators were working. Investigators did have shifts including days, evenings and nights.  All patients 18 years and older, with and ESI designation 2-4 were included. Exclusion criteria included those with ESI designations 1 and 5, or if patients refused to consent for the study. Enrollment period was from February 2016 to August 2018.

Study protocol involved investigators measuring EtCO2 of patients arriving via the lobby or EMS. Investigators included a range of training levels from medical students to residents and board certified Emergency Medicine physicians. Investigators underwent a 1 hour training session. Measurements were recording by taking the EtCO2 value on the 8th breath of patients. If a patient already had EtCO2 measured prior to the investigator, they used that value. All nasal cannulas were the same brand, including those in the lobby and used by EMS. The treatment team was blinded by these values.  

The primary outcome was in-hospital mortality. Patients discharged home, including from the ED, were regarded as having survived to hospital discharge. The secondary outcome focused on ICU admissions. The tertiary outcome was the relationship between EtCO2 and serum lactate, anion gap and bicarb.

Results

There were 1136 patients prospectively enrolled in the study and 1091 patients with both EtCO2 and outcome data available. Patient’s mean age was 56 years and 53% were male. The general characteristics between survivors and non-survivors were comparable with the exception of age and mental status. Patients who survived were significantly younger (55 years vs. 68 years) and had a higher proportion of normal mental status (88% vs. 35%). There also were significant differences in race, with a higher proportion of Hispanics and African Americans among survivors and more “unknown” and “other” races in non-survivors. Hospital admission and ICU admission were both significantly higher in non-survivors.

The mean level EtCO2 in all patients was 34 (95% CI 33-34) mmHg.

There were 26 (2.4%) patients with in-hospital mortality and mean EtCO2 levels in survivors vs. non-survivors were 34 (95% CI 33-34) vs. 22 (95% CI 18-26, p < 0.001).

The area under the curve (AUC) for predicting in-hospital mortality for EtCO2 was 0.82 (0.72—0.91), AUC for temp was 0.55 (0.42-0.68), AUC for RR was 0.59 (0.46-0.73), AUC for systolic blood pressure (SBP) was 0.77 (0.67-0.86), AUC for diastolic blood pressure (DBP) was 0.70 (0.59-0.81), AUC for HR was 0.76 (0.66-0.85), and AUC for SpO2 was 0.53 (0.40-0.67). In adjusting for all vital signs, EtCO2 remained the strongest predictor of mortality 0.88 (95% CI 0.83-0.93, p < 0.001) followed by pulse 1.023 (95% CI 1.005-1.041 p=0.013) and SBP 0.97 (95% CI 0.95-1.00 p=0.051). There were 64 (6%) patients admitted to the ICU, and the EtCO2 AUC for predicting ICU admissions was 0.75 (0.67-0.80), temperature 0.51 (0.42–0.59), RR 0.56 (0.47–0.65), SBP 0.64 (0.56–0.72), DBP 0.63 (0.55–0.71), HR 0.66 (0.58–0.73), and SpO2 0.53 (0.45–0.61). There were significant correlations between expired EtCO2 and 3 serum lab measures of metabolic acidosis including serum lactate, anion gap and sodium bicarbonate, with rho = -0.26 (p <0.001), rho = -0.20 ( p < 0.001) and rho = 0.330 (p < 0.001), respectively.

They did explore cutoff points of ETCO2 for predicting in-hospital mortality using the AUROC curve to maximize sensitivity and specificity. Using a cutoff of 28 mm Hg for ETCO2 yielded a sensitivity of 77% (95% CI 56%–90%) and a specificity of 79% (95% CI 77%–82%) with a negative predictive value of 99% (95% CI 98%–100%), a positive predictive value of 8% (95% CI 5%–13%).

Limitations

While an interesting study that attempts to address a growing issue in the field of Emergency Medicine, there are several limitations to this study. This study did use a convenience sample, which may have led to sample bias. The time period of 30 months helps mitigate this somewhat, with a variety of clinical shifts from the investigators, but nonetheless this can’t be ignored. This was performed at a single ED. One of the bigger limitations was the use of a single data point of EtCO2 rather than a continuous measurement. The utility of EtCO2 in most settings relies on the trend and continuous nature of the variable rather than a single point in time. You can make an argument that could apply to blood pressure and HR in critically ill patients, however this is still a major limitation as trend may be more useful than a singular data point. The authors highlighted inability to obtain the race of many of the non-survivors may have affected the proportion of Hispanics and African Americans in the survivor group. One of the limitations we had discussed in Journal Club was the fact this study measured EtCO2 against each individual vital sign rather than as a collective set. An additional limitation was including “survived to discharge” as those who were discharged from the ED, as they may have gone to another hospital and either been admitted or potentially died shortly after.

Takeaway

I picked this study because it attempts to address the issue of how to identify sick patients sitting in the lobby for long periods of time in the context of overcrowding and boarding. As our systems continue to be constrained, our triage process is going to need to adapt to efficiently re-evaluate patients on an interim and accurately identify those patients who are critically ill or may become so. EtCO2 is an intriguing measurement that has its roots in physiology, making it useful as an additional data point in triaging sick patients. Whether it’s superior to our current vital signs is questionable and requires further studies. It’s use as a continuous variable is much more valuable, however if we can identify a cutoff point after which we intervene, it may become more useful as a single data point. The most useful purpose would be to use EtCO2 to replace the respiratory rate, as this is highly inaccurate. Having a more reliable method to identify tachypnea may be a better use of EtCO2. Cost is an additional limiting factor, as disposable nasal cannulas are not cheap. All in all, it is an excellent study to spark a conversation regarding how to best identify sick patients early in the course of their disease as well as how to best utilize EtCO2, whether that be as a continuous variable, singular data point with an acceptable cutoff point, or utilizing the EtCO2 to supplement a more accurate RR. This requires further study before it is ready for prime time.

References

  1. Edmonds ZV, Mower WR, Lovato LM, Lomeli R. The reliability of vital sign measurements. Ann Emerg Med. 2002;39(3):233-237.


  2. Ward KR, Yealy DM. End-tidal carbon dioxide monitoring in emergency medicine, part 1: basic principles. Acad Emerg Med. 1998;5(6):628-636.


  3. Ward KR, Yealy DM. End-tidal carbon dioxide monitoring in emergency medicine, part 2: clinical applications. Acad Emerg Med. 1998;5(6):637-646.


  4. Hunter CL, Silvestri S, Ralls G, et al. Comparing quick sequential organ failure assessment scores to end-tidal carbon dioxide as mor- tality predictors in prehospital patients with suspected sepsis. West J Emerg Med. 2018;19(3):446-451. 


  5. Hunter CL, Silvestri S, Ralls G, Bright S, Papa L. The sixth vital sign: prehospital end-tidal carbon dioxide predicts in- hospital mortality and metabolic disturbances. Am J Emerg Med. 2014;32(2):160-165. 


  6. Ladde JG, Miller S, Chin K, et al. End‐Tidal Carbon Dioxide Measured at Emergency Department Triage Outperforms Standard Triage Vital Signs in Predicting In‐Hospital Mortality and ICU admission . Acad Emerg Med. 2023;(December 2022):1-10. doi:10.1111/acem.14703

Authorship

Written by and Podcast Audio by Anthony Martella, PGY-3, University of Cincinnati Department of Emergency Medicine

Peer Review, Editing, Audio Editing, and Posting by Jeffery Hill, MD MEd, Associate Professor, University of Cincinnati Department of Emergency Medicine

Cite As:

Martella, A. Hill, J. EtCO2 vs. Standard Triage Vitals in Predicting In-Hospital Mortality and ICU Admission. TamingtheSRU. www.tamingthesru.com/blog/journal-club/etco2-vs-standard-triage-vitals-in-predicting-in-hospital-mortality-and-icu-admission. 6/11/2023