Prove it: Using capnography to guide ventilation rates
There is a growing body of evidence that suggests that hyperventilation of intubated patients with head injuries increases both morbidity and mortality
Rescue 3 responds to a reported traffic collision in the early morning hours. Arriving on the scene, the medics find a small car that left the roadway, overturned, and rolled several times. The only patient is an unconscious male who looks to be in his mid-20s. He is lying in the ditch about 20 feet from the upside-down vehicle. The patient is making agonal respiratory efforts.
One firefighter holds the patient's head in a neutral position while one paramedic successfully performs an in-line endotracheal intubation (confirmed with waveform capnography) and secures the tube. The patient's radial and brachial pulses are absent, but the medic reports a weak carotid pulse at 120 beats per minute.
A rapid head-to-toe exam reveals a right femur fracture, a laceration across the forehead, and extensive abrasions and bruising across the anterior chest. The patient's pupils are reacting sluggishly to the light. Concerned about the potential for a traumatic brain injury, the medic attempts to reduce brain swelling by having the firefighter ventilate the patient at a rate necessary to keep the capnography reading near but not less than 35 mmHg.
There is a growing body of evidence that suggests that hyperventilation of intubated head injured patients increases both morbidity and mortality (Davis, et al, 2004; Davis, et al, 2006; Warner, Cuschieri, Copass, Jurkovich, & Bulger, 2008). As a result, emergency medical service administrators, educators, and medical directors stress the importance of maintaining capnography levels within normal limits when assisting ventilation in these patients.
The International Trauma Life Support (ITLS) guidelines recommend that EMS providers should ventilate head injured patients at a rate necessary to keep the end-tidal carbon dioxide (EtCO2) levels between 35 and 40 mm Hg (Alson, & Campbell, 2008). Hyperventilation (EtCO2 less than 35 mmHg) is appropriate only when the patient has evidence of cerebral herniation (Badjatia, et al, 2008).
The question, however, is whether EMS can rely on end-tidal capnography readings to estimate arterial carbon dioxide levels within the patient's blood stream.
In a prospective observation study conducted in Washington State, researchers examined the relationship between EtCO2 and the partial pressure of carbon dioxide found in arterial blood (PaCO2) (Warner, Cuschieri, Garland, Carlborn, Baker, Copass, et al., 2009).
In a prospective investigation, researchers first design the study and subsequently collect data as it becomes available. In contrast, data in a retrospective study already exists (such as in a chart or database) before the researchers design the study. Prospective studies offer researchers the advantage of being able to control more variables. For example, after designing this study, the researchers identified the cohort (group) of test subjects and began collecting data on each patient as he or she arrived in the emergency department. The research team had better control over whether each test subject received the tests needed for inclusion into the study group.
The population (cohort) for this study was a group of 180 consecutive trauma patients intubated by prehospital personnel prior to arrival at a single Level 1 trauma center. The cohort did not include patients less than 18 years of age, primary burn patients, or those triaged to a resuscitation bed that lacked the ability to monitor EtCO2 values. In some cases, emergency department staff moved patients to surgical suites before implementing the study protocol and those were not included in this group.
As expected in any cohort of trauma patients, most (77 percent) were young males (median age 34). Only 9 percent of the patients suffered exclusively from penetrating injury. The majority (61 percent) suffered a severe brain injury (severe was defined as an abbreviated injury score [AIS] of ≥ 3), almost half (46 percent) suffered a severe chest injury (AIS ≥ 3) and over one-fourth (28 percent) had impaired tissue perfusion (arterial base deficit > 6 mEq/L).
Once in the ED, respiratory therapists (RT) placed each patient on a mechanical ventilator along with continuous end-tidal capnography. Researchers instructed each RT to determine ventilator settings based on the patient's estimated ideal body weight and not to make adjustments based on EtCO2 levels. During arterial blood gas collection, ED personnel recorded the EtCO2 reading at the moment of the arterial puncture, make those simultaneous measurements. Statisticians used linear regression to determine the correlation between EtCO2 levels and PaCO2 values.
Researchers demonstrated a statistically significant relationship between EtCO2 and PaCO2 (p < 0.001). This means, in general, that when EtCO2 values were low, PaCO2 levels were low. When EtCO2 was high, PaCO2 was high. For any given value of EtCO2, the corresponding value of PaCO2 was, on average, about 8 mm Hg higher.
However, in 95 percent of the individual patients (confidence interval (CI)), the PaCO2 value fell in a range from 8 mm Hg lower to 25 mm Hg higher than the corresponding EtCO2 (95 percent CI [-8, +25]). Obviously, then, 5 percent of the patients had PaCO2 values even farther from the corresponding EtCO2 levels. The authors concluded that because of this wide confidence interval, the predictive value of EtCO2 readings on determining PaCO2 levels in trauma victims is questionable.
They add that if paramedics had implemented the ITLS guideline recommendation of ventilation to maintain an EtCO2 reading between 35 and 40 mm Hg in this cohort of trauma patients, 80 percent would have PaCO2 levels greater than 40 mm Hg and 30 percent would have PaCO2 levels greater than 50 mm Hg, potentially placing these patients at greater risk of adverse outcomes.
The authors conclude that there appears to be poor correlation between EtCO2 and PaCO2 readings in trauma patients. It is likely that EtCO2 readings in trauma victims, especially multi-system victims, are more accurate measures of perfusion than of ventilation status. Additionally, the EtCO2 readings in patients with chest injury may be reflecting a ventilation-perfusion mismatch.
Several limitations to this study influence how widely EMS personnel can apply the results. First, the study suffers from a selection bias, meaning that the process of selecting the study population created a group with different characteristics than the population at large. The study authors drew conclusions about intubated trauma patients in general.
However, the researchers could not collect data on about one-fifth of all the intubated trauma patients arriving at the trauma center because they went directly to the operating room without stopping in the emergency department. Researchers excluded another 33 percent because the triage team did not send those eligible patients to the resuscitation bay where the trauma team collected the study data. In fact, only 31 percent of all the eligible study participants had enough data for inclusion into the investigation.
Additionally, this small subset of intubated trauma patients suffered a variety of injury patterns making the group heterogeneous, or different. The only thing the group had in common was that paramedics intubated them in the field and they suffered some type of trauma. However, some had isolated head injuries, some had isolated chest injuries, some were hemorrhaging, and many had a combination of injuries.
The highest correlation between EtCO2 and PaCO2 readings was in the isolated head injury group while the worst was in the abdominal injury group and those with a low shock index (heart rate divided by systolic blood pressure).
Unfortunately, the study does not provide enough information about any single injury pattern to draw meaningful conclusions or to explain the inconsistencies adequately.
Many EMS systems may already have or are contemplating the adoption of EMS treatment protocols based on the ITLS recommendation of using EtCO2 as a guide for ventilation rates in some patients, such as intubated trauma patients. There is little question that capnography has value within an EMS system, especially for advanced airway placement confirmation and tube surveillance. However, using EtCO2 readings to guide ventilation rate in all trauma patients may provide advantages for some, but disadvantages for others. Further research must help distinguish between the two.
Alson, R. L., & Campbell, J E. (2008). Head trauma. In J. E. Campbell (Ed.), International trauma life support for prehospital care providers (6th ed., pp. 141-159). Upper Saddle River, NJ: Pearson Education.
Badjatia. N., Carney, N., Crocco, T. J., Fallat, M. E., Hennes, H. M., Jagoda, A. S., et al. (2008). Guidelines for prehospital management of traumatic brain injury (2nd ed.). Prehospital Emergency Care, 12 Suppl 1: S1-S52.
Davis, D. P., Dunford, J. V., Poste, J. C., Ochs, M., Holbrook, T., Fortlage, D., et al. (2004). The impact of hypoxia and hyperventilation on outcome after paramedic rapid sequence intubation of severely head-injured patients. The Journal of Trauma, 57, 1– 8.
Davis, D. P., Idris, A. H., Sise, M. J., Kennedy, F., Eastman, A. B., Velky, T., et al. (2006). Early ventilation and outcome in patients with moderate to severe traumatic brain injury. Critical Care Medicine, 34, 1202–1208.
Warner, K. J., Cuschieri, J., Copass, M. K., Jurkovich, G. J., & Bulger, E. M. (2008). Emergency department ventilation effects outcome in severe traumatic brain injury. The Journal of Trauma, 64, 341–347.
Warner, K. J., Cuschieri, J., Garland, B., Carlborn, D., Baker, D., Copass, M. K., et al. (2009). The utility of early end-tidal capnography in monitoring ventilation status after severe injury. The Journal of Trauma, 66, 26 –31.
The author has no financial interest, arrangement, or direct affiliation with any corporation that has a direct interest in the subject matter of this presentation, including manufacturer(s) of any products or provider(s) of services mentioned.