Caution is advised when attempting to use EtCO2 to approximate an arterial blood gas CO2 (PaCO2). While EtCO2 can correlate with PaCO2 within 5 mm Hg in greater than 80% of patients with dyspnea,44 large discrepancies are common depending on the disease state.45 In general, the EtCO2 should always be lower than the PaCO2 due to the contribution to the ETCO2 from dead space, which has a low CO2 content due to lack of perfusion.
Sepsis
EtCO2 may help identify septic patients given its inverse relationship with lactate levels.46-49 In conditions of poor tissue perfusion, lactate builds up. This begins to make the blood acidotic in the form of newly acquired anions, with a resultant anion gap metabolic acidosis. The body then tries to acutely compensate for this by hyperventilating, resulting in the observed lowering of EtCO2. Since lactate is a predictor of mortality in sepsis,50 and monitoring lactate clearance to evaluate resuscitation efforts in sepsis is recommended,51 EtCO2 could play a similar role. One group in particular has demonstrated that, when used with SIRS criteria, abnormally low prehospital EtCO2 levels is predictive of sepsis and inhospital mortality, and is more predictive than SIRS criteria alone.48,50 That said, EtCO2 was not associated with lactate temporally at 3 and 6 hours,51 so it should not be used to guide resuscitation like a lactate clearance. It appears that EtCO2 may be helpful for triage in sepsis, but more study is needed to determine the exact role particularly given most of the available research involves multiple studies from one group.47,48,52
Diabetic Ketoacidosis
Initial bicarbonate levels and venous pH are associated with low EtCO2 readings in diabetic ketoacidosis (DKA).54,55 This could have many practical uses, in particular for patients presenting with hyperglycemia to rule out DKA. One study demonstrated that a blood glucose >250 mg/dL and capnography of >24.5 mm Hg had 90% sensitivity for excluding DKA.55 A value of 35 mm Hg or greater demonstrated 100% sensitivity for excluding DKA in patients with initial glucose >550 mg/dL,56 though this blood glucose is not practical, as this excludes many patients the EP would seek to rule out DKA (recall that blood glucose only has to be >250 mg/dL for the diagnosis). Smaller studies focused on the pediatric population found a 100% sensitivity marker for DKA varied from >30 to >36 mm Hg.57,58 Clearly a role exists, but no study has demonstrated sufficient sensitivity for ruling out DKA with EtCO2 and blood glucose alone within the framework of clinically relevant values.
Trauma
As described above, low EtCO2 is inversely correlated with lactate.46 Because of this, it could theoretically be a marker of hypoperfusion in trauma. Initial EtCO2 values <25 mm Hg have been associated with mortality and hemorrhage in intubated trauma patients,59 as well as mortality prior to discharge in nonintubated trauma patients.60 However, it did not demonstrate added clinical utility when combined with Glasgow Coma Scale (GCS) score, systolic blood pressure, and age in predicting severe injury.61
Pulmonary Embolism
A pulmonary embolism (PE) causes a blockage in blood flow to alveoli, which results in a decrease in CO2 transportation to the alveoli and thus lower EtCO2, while also widening the gradient between PaCO2 and EtCO2.37 Because of this, it has a theoretical role in the diagnosis of PE, though numerous studies have demonstrated that EtCO2 alone is not sensitive nor specific enough for this role.62-66 In a recent meta-analysis, a pretest probability of 10% could lead to a posttest probability of 3% using capnography.62 While further study is needed before recommendation, this indicates that capnography could obviate the need for imaging in low to intermediate risk patients either after a positive D-Dimer or instead of obtaining a D-dimer.62-64