Critical Care Alert: ETCO2 as a Predictor of Patient Outcomes

Article: Portelli Tremont JN, Caldas RA, Cook N, Udekwu PO, Moore SM. Low initial in-hospital end-tidal carbon dioxide predicts poor patient outcomes and is a useful trauma bay adjunct. Am J Emerg Med. 2022;56:45-50. 

Objective: To evaluate the association between end-tidal capnography (ETCO2) in non-intubated trauma patients and blood product transfusion product requirements (primary outcome), other clinical outcomes, and to extrapolate the utility of ETCO2 in early trauma management.  

Background: The triage and determination of patients in shock in the trauma setting continues to be difficult and has limited parameters. End tidal capnography (ETCO2) is commonly utilized by clinicians to access a critical patient’s hemodynamic and metabolic fluctuations, as rising lactate levels with tissue malperfusion are generally accompanied by a decrease in ETCO2.^1,2 While commonly used to evaluate the quality of chest compressions and appropriate placement of advanced airway, ETCO2 has not been robustly studied in the triage setting.^3-5

Furthermore, prior studies focused on mainstream capnography in intubated patients, but a recent retrospective cohort study of non-intubated trauma patients who received pre-hospital side-stream capnography showed that low ETCO2 was associated with increased morbidity and mortality.^1,4  

As such, this article is of a single-center retrospective observational cohort study that aims to show that non-invasive (side-stream) ETCO2 monitoring in early trauma resuscitation can be useful in assessing occult shock  

Inclusion Criteria

• Age >16 years old
• Initial ETCO2 obtained from side-stream nasal canula capnography
• Patients with measured trauma bay ETCO2
• Patient presenting as a trauma activation or having been entered into the trauma registry between 2019 and 2020

Exclusion Criteria

• Patient intubated prior to ETCO2 measurement
• Patients with erroneous ETCO2 (0- and 1mmHg) values

METHODS
This study was a single-center retrospective observational cohort study of 955 patients that analyzed the association between initial in-hospital ETCO2 of non-intubated trauma patients and transfusion requirements, as well as other clinical outcomes.  

Patients were dichotomized into low (<29.5 mmHg) or normal/high (>29.5 mmHg) ETCO2 groups by the Lui method of cut point estimation which maximizes specificity and sensitivity.

Patient demographics and clinical characteristics were compared using Chi-square test for categorical data and Wilcoxon rank some tests for continuous variables. 

Unadjusted and multivariable logistic and linear regression were used to estimate the association between ETCO2 and the primary and secondary outcomes.  

ETCO2 was also compared to the validated shock index (SI) to predict the need for blood transfusions (>4 units in the first 24 hours). This comparison was made by calculated sensitivity, specificity, and area under the receiver operating curve (ROC) for ETCO2 and SI. 

Primary Outcome

• Transfusion of blood products (>4 units) with 24 hrs of presentation

Secondary Outcomes

• Total hospital length of stay (LOS)
• All-cause mortality (including deaths occurring in the trauma bay)
• ICU admission
• Any hospital complication (pneumonia, sepsis, wound infection, alcohol withdrawal, pulmonary embolism, deep vein thrombosis, respiratory failure)
• Need for mechanical ventilation
• Discharge disposition

RESULTS
Demographics

• Median collection time of non-invasive side-stream ETCO2 was within 4 minutes of arrival
• Median ETCO2 in the low ETCO2 group is 22mmHg, and 35mmHg in the normal/high ETCO2
• Patients with lower ETCO2 were significantly older than those with normal/high ETCO2 (53 vs 45, p=0.011)
• Injury mechanism and vital signs were statistically similar between both groups (p >0.05) - This includes mostly vehicle collision, fall, assault, and penetrating trauma  
• Patients with low ETCO2 were likely to have higher trauma activation (27.4% vs 19.8%, p=0.048)

Blood transfusion, mortality, and other clinical outcomes

• Patients with low ETCO2 are 4.65 times more likely to needed blood transfusions, compared to 3.6% of the patients with normal/high ETCO2 (95% CI 2.01, 10.72 P<0.001)
• Patients with low ETCO2 are 5.10 times greater odds of mortality than those with normal/high ETCO2 (95% CI 1.05-24.90, p=0.044)
• Regarding other clinical outcomes, patients with low ETCO2 were significantly more likely to require mechanical ventilation, emergent operation, experience inpatient complications, have longer hospital length of stay, and be discharged to a skilled nursing facility.
• No significant difference between the groups regarding outcomes among those with head injury, but patients with head injury and low ETCO2 have decreased likelihood of hospital admission

ETCO2 compared to Shock Index (SI)

• ETCO2 has a greater sensitivity (77.8% vs. 33.3%) but less specificity (57.8% vs 95.0%) of predicting need for blood transfusion than a calculated SI

CONCLUSIONS

• Low ETCO2 is significantly associated with poor outcomes in trauma patients compared to those with normal/high ETCO2
• ETCO2 is a more sensitive tool at predicting need for blood transfusion than SI in the acute trauma setting.

Limitations

• Choice of patients to receive ETCO2 monitoring was at the discretion of the trauma attending.
• Patients may have received pre-hospital medications that can affect ETCO2
• Non-invasive ETCO2 more likely to be affected by patient factors than mainstream ETCO2, such as mouth breathing, hyperventilation, excess secretions.
• The dichotomy of ETCO2 limits the ability to analyze as only two sets as opposed to examining trends over various ranges of ETCO2.  
• The cut point used is not based on physiology and based on the patient population. 

EM TAKE-AWAYS
The triage of patients in the trauma bay is difficult and no single measurement or vital sign will provide sufficient information for mortality. The job of the trauma team, including the EM physician, is to utilize many factors and this paper supports the addition of ETCO2 as one of those markers. ETCO2 is a commonly accessible tool in the ED and can be utilized to raise awareness that a patient may have increased odds of mortality or need for blood transfusion based on a lower value (22mmHg vs 35mmHg) 

• Non-invasive ETCO2 monitoring is more commonly accessible and related to the emergency department setting
• ETCO2 provides a more readily accessible compared to other parameters of assessment.

REFERENCES

1. L. Hunter, S. Silvestri, M. Dean, J.L. Falk, L. Papa. End-tidal carbon dioxide is associated with mortality and lactate in patients with suspected sepsis. Am J Emerg Med, 31 (1) (2013), pp. 64-71
2. E. Stone Jr., S. Kalata, A. Liveris, Z. Adorno, S. Yellin, E. Chao, et al. End-tidal CO2 on admission is associated with hemorrhagic shock and predicts the need for massive transfusion as defined by the critical administration threshold: a pilot study. Injury., 48 (1) (2017), pp. 51-57
3. Long, A. Koyfman, M.A. Vivirito. Capnography in the emergency department: a review of uses, waveforms, and limitations. J Emerg Med, 53 (6) (2017), pp. 829-842
4. J. Williams, F.W. Guirgis, T.K. Morrissey, Wears R.L. Wilkerson, C. Kalynych, A.J. Kerwin, et al. End-tidal carbon dioxide and occult injury in trauma patients: ETCO2 does not rule out severe injury. Am J Emerg Med, 34 (11) (2016), pp. 2146-2149
5. Pantazopoulos, T. Xanthos, I. Pantazopoulos, A. Papalois, E. Kouskouni, N. Iacovidou. A review of carbon dioxide monitoring during adult cardiopulmonary resuscitation. Heart Lung Circ, 24 (11) (2015), pp. 1053-1061
6. Bryant, M. K., Tremont, J. N. P., Patel, Z., Cook, N., Udekwu, P., Reid, T., ... & Moore, S. M. Low initial pre-hospital end-tidal carbon dioxide predicts inferior clinical outcomes in trauma patients. Injury, 52(9) (2021)., pp. 2502-2507.