end tidal CO2, what intraoperative role can it play?

Today I had a 64 year old male, with no reported medical history, who presented for L4-5 laminectomy.  The patient reported that he walked regularly, up to 2 miles with no history of SOB, chest pain or other symptoms.  The patient was taken to the OR and given 2 mg versed, 100 mcg fentanyl, 180 mg propofol, 5 mg rocuronium, and 100 mg succinylcholine.  IV decadron was also given (8mg).  Intubation proceeded without event and the patient was placed in the prone position.  After turning prone, several issues arose at the same time.  First, pulse oximetry revealed 93% saturation.  Simultaneously, the blood pressure read 50/20 mmHg. ECG appeared normal with SR at 74 bpm.  The pulse plethysmograph waveform appeared robust without obvious issues aberrations.
Ausculation of the lungs was challenging due to a large tissue mass making breath sounds difficult to detect.  However, it was noted by myself, that there appeared to be no breath sounds on the left, and therefore, the ETT was pulled back slightly.  This resulted in an improvement of the arterial saturation as measured by pulse oximetry.  However, simultaneous troubleshooting of the significant apparent hypotension occurred.  The blood pressure cuff was recycled, and low blood pressure was verified.  Also of note, the capnogram was reading 20 mmHg.  The patient was noted to be ventilated at 700 mL with RR of 10.  The patient weighed approximately 100 to 105 kg. It was immediately apparent that elevated minute ventilation was not likely the sole contributor to the issues related to the hypocapnea.

While multiple issues were at play at once in this case (endobronchial intubation along with severe hypotension), a deeper exploration of how the capnogram can be helpful in the diagnosis of the issues  at hand.

In general, end tidal carbon dioxide (etCO2), is a function of PaCO2.  However, a multitude of parameters can cause a gap (this is dead space [Vd], written as (a-ET)PCO2).  In general, healthy patients without significant lung disease will have up to a 5 mmHg difference, where the etCO2 will be less than actual measure PaCO2.    This gap increases with age, emphysema or any state that increass dead space ventilation (Vd), like low cardiac output (from hypovolemia) or pulmonary embolism.  On the other hand, (a-ET)PCO2 can actually be positive (i.e. etCO2 is greater than PaCO2) in pregnancy and children (from 1 to 3 mmHg). In general clinical practice, we do not have access to the (a-ET)PCO2 because we do not routinely measure arterial blood gases.  However, we do follow the trend of the etCO2, and thus, in general, if we see a sudden decrease in the etCO2 on the capnogram, we assume that we may be hyperventilating the patient, or that there has been a sudden increase in dead space ventilation.  However, it must be understood, that there are several parameters other than Vd that can cause a decrease in etCO2. For example, decrease in metabolism or VCO2 will result in decrease in etCO2 if minute ventilation is constant. VCO2 is a function of depth of anesthesia relative to surgical stimulus, and body temperature.  Alternatively, an increase in minute ventilation, if metabolic rate is constant will cause etCO2 to decrease.  Importantly, the (a-ET)PC02, will remain the same in the two above situations.  Another, more sutbtle and less recognized mechanism for (a-ET)PCO2 to be affected is via FiO2.  Yamauchi et al. demonstrated that increasing the FiO2 from room air to very high caused an ever increasing (a-ET)PCO2. [8] They found that Vd increased as the FiO2 was increased in their anesthetized patients.  They presumed that the mechanism of the increase in Vd was an increase in pulmonary vascular dilation with increased oxygen tension.  This occurs predominantly in highly perfused alveolar units resulting in a shunt of blood away from low perfused alveolar units to high perfused alveolar units.  Of course, this shunt created from increasing FiO2 is not related to physiologic shunts that might occur with something like ARDS.  In this case, only large shunts of greater than 30 to 40% will cause a significant change in the (a-ET)PCO2.

However, a state of low cardiac output can also result in a reduced pulmonary artery blood flow. This results in increased Vd, and thus the (a-ET)PC02 increases.  This manifests in the operating room as a sudden decrease of etCO2.  This pattern was looked at by Askrog and colleagues where an inverse linear correlation was found between pulmonary artery pressure and (a-ET)PCO2. [1]     Things that can cause this include pulmonary emboli (air, debris, clots), sudden massive hemorrhage leading to reduced venous blood return, vasodilation, mechanical obstruction to blood flow, reduced cardiac contractility, etc.  In general, in the OR, mechanical ventilation and anesthetic depth are maintained at a reasonably constant level allowing us to remove these as a cause in theory.

In my patient, the simultaneous low blood pressure and sudden drop in etCO2, indicated two things: the blood pressure was real (i.e. it was not artifactual), and the drop in etCO2 was most likely due to decreased CO, in this case  the cause being overdose of anesthesia. Of course other causes of a precipitous drop in etCO2 include PE or other mechanical obstruction to pulmonary blood flow.  As it turns out the percent decrease in etCO2 is directly correlated with the percent decrease in CO (assuming that the metabolic rate and alveolar ventilation are unchanged). This was demonstrated in an article published in A and A. [2]  It should be noted that after sustained or constant low CO, (i.e. after 10 to 20 min) CO2 begins to accumulate in the peripheral tissues leading to an increase in CO2 delivery to the pulmonary vasculature.  This will cause the etCO2 to return to baseline if all other factors remain unchanged.  The relationship of etCO2 and pulmonary blood flow was also studied in patients coming off cardiopulmonary bypass. [3]  Here, an etCO2 greater than 30 mmHg (the study did not include patients with significant lung disease), was associated with CO of greater than 4 L/min (CI of 2 L/m). When etCO2 was greater than 34 mmHg, pulmonary blood flow (a good surrogate for CO) was greater than 5 L/min. Once again, minute ventilation was carefully maintained.
Recently, a large volume of literature has been produced looking at measurements of indices that indicate a patient who is hypovolemic. Pulse pressure variation and stroke volume variation via measurement and analysis of the arterial waveform in ventilated patients in sinus rhythm has proven effective at determing which patients are likely going to respond with increased cardiac output if a fluid bolus is given.  Unfortunately, the equipment is costly, requires a fair amount of data input, and is usually not routinely available. Recently, Monnet et al. [4] was able to show that etCO2 was better than arterial pressure for predicting volume responsiveness when using a passive leg raising test. Using a similar methodology in patients with acute circulatory failure in the ICU, monge garcia et al. showed that etCO2 after passive leg raise maneuver could be used to track changes in CO  for the prediction of fluid responsiveness. [5]  Recently, a group in France was able to demonstrate that after 500 mL hetastarch, an increasae of 2 mmHg in the etCO2 could diagnose fluid responsiveness (specificity 98%, sensitivity 60%). Obviously, it is critical to undertand that other changes to CO2 production and elimination must be held constant for this to be a valid clinical indicator.  It should be recognized that in clinical anesthesia this can be difficult.  In fact, very recently, I took care of patient having an open partial colectomy with small bowell resection.  Her BP trailed lower early in the case.  Based on this article, I decided to carefully track etCO2 and maintain other parameters unchanged (i.e. CO2 production and minute ventilation).  I quickly boluses in 500 mL of hetastarch as used in the above mentioned article.  While I did notice that etCO2 trailed higher with this bolus, I realized that in clinical practice, there are so many other factors occurring that it can be challenging to feel confident that other parameters are not the cause of the change you see reading on the capnogram.   However, importantly,  in this same study, HR variation, MAP variation, and PP variation were not predictive of volume responsiveness. In summary, these authors showed that after a rapid infusion of 500 mL colloid, an increase of 5.8% (or about 2 mmHg) of etCO2 predicted fluid responsiveness in 100% of their patients.  If the etCO2 increased less than 5.8%, no conclusions could be drawn. [6]  In real clinical practice, a bolus must be given very rapidly, to ensure that other parameters don't account for any subtle changes seen on the etCO2.
Others have demonstrated that etCO2 can be predictive of mortality after out of hospital cardiac arrest (OHCA). In the NEJM [7], an observational study was published looking at etCO2 monitoring following OHCA to determine effectiveness of ACLS.  They found that in this patient population, if  after 20 minutes of ACLS, the etCO2 was less than 10 mmHg, there was a 100% specificity and specificity to determine non survival to hospital admission.  If the etCO2 was greater than 20 mmHg, this indicates survival (at least in this study), but it does not guarantee it.  In this article, it was noted that in low flow states (i.e. low cardiac output), etCO2 becomes a much better surrogate for cardiac output.

Having an in depth understanding of etCO2 can help us in ways that we might not otherwise expect.





[1]  Askrog V.  Changes in  (a-A)CO2 difference and pulmonary artery pressure in anesthetized man. J Appl Physiol 1966;;21:1299-1305.
[2] Shibutani K, Muraoka M, Shirasaki S, Kabul K, Sanchala VT, Gupte P.  Do changes in end-tidal PCO2 quantitatively reflect changes in cardiac output? Anesth Analg 1994;79:829-33.
[3] Maslow A, Stearns G, Bert A, Feng W, Price D, Schwartz C, Mackinnon S, Rotenberg F, Hopkins R, Cooper G, Singh A, Loring SH. Monitoring end-tidal carbon dioxide during weaning from cardiopulmonary bypass in patients without significant lung disease. Anesth Analg 2001;92:306-13.
[4] Monnet X, Bataille A, Magalhaes E, Barrois J, Le Corre M, Gosset C, Guerin L, Richard C, Teboul J-L. End-tidal carbon dioxide is better than arterial pressure for predicting volume responsiveness by the passive leg raising test. Intensive Care Med 2013;39: 93–100. 
[5] Monge García MI, Gil Cano A, Gracia Romero M, Monterroso Pintado R, Pérez Madueño V, Díaz Monrové JC. Non-invasive assessment of fluid responsiveness by changes in partial end-tidal CO2 pressure during a passive leg-raising maneuver. Ann Intensive Care. 2012;2:9
[6] Jacquet-Lagreze M, baudin F, David JS, Fellahi JL, Hu PB, Lilot Ma, Piriou V. End -tidal carbond dioxide variation after a 100- and a 500-ml fluid challenge to assess fluid responsiveness.  Annals of Intensive Care 2016 6:37.
[7]Levine R, et al. End Tidal Carbon Dioxide and Outcome of Out-of-Hospital Cardiac Arrest. NEJM; 337(5):301.
[8] Yamauchi H, Ito S, Sasano H, Azami T, Fisher T, Sobue K.  Dependence of the gradient between arterial and end-tidal PCO2 on the fraction of inspired oxygen.  BJA. 2011.107(4):631-5.