ICU Management & Practice, Volume 18 - Issue 4, 2018

Advances in monitoring expired CO2 in critically ill patients

share Share
Reviews the potential uses and pitfalls of capnography in critically ill patients, especially for haemodynamic and respiratory monitoring.

Expired CO2 can be easily monitored in the intensive care unit (ICU), especially in patients under invasive mechanical ventilation, using infrared measurement by sampling mainstream expiratory flow using an in-line chamber, or sidestream expiratory flow (by continuous aspiration through a sampling line connected between the intubation tube and the Y-piece of the ventilator).


Expired CO2 is determined by three parameters: 

  1. CO2 production (CO2) mainly due to tissue metabolic activity
  2. CO2 transport related to cardiac output (CO) and haemoglobin level
  3. CO2 clearance by alveolar ventilation. 

Given its high diffusive capacity, CO2 is easily eliminated by alveolar ventilation, although end-tidal CO2 partial pressure (PEtCO2) is higher than alveolar CO2 partial pressure (PACO2) due to ventilation-perfusion mismatch. The gradient between PEtCO2 and arterial CO2 partial pressure (PaCO2) is usually low (3-5 mmHg) but increases with increasing alveolar dead space, even though PEtCO2 remains highly correlated with PaCO2.


By allowing a combined analysis of respiratory, haemodynamic and metabolic status, capnography is a versatile tool with developing clinical applications in the ICU. While capnography is commonly used in the operating room, this technique may be underused in the ICU (Cook et al. 2011; Georgiou et al. 2010; Ono et al. 2016).The aim of this paper is to highlight the advances and the usefulness of expired CO2 monitoring in the specific ICU setting.


Capnography and respiratory intensive care


Airway management

Capnography is a reliable tool to confirm the correct placement of endotracheal (Guggenberger et al. 1989) or supraglottic devices, given the lack of significant CO2 production in the oesophagus. Despite a high diagnostic performance (Silvestri et al. 2005), false negatives are encountered in the cardiac arrest setting (Heradstveit et al. 2012) or as a consequence of technical pitfalls (leaks around endotracheal cuff (Dunn et al. 1990), kinking of the sampling line, ventilator failure…). False positives may happen if the stomach contains CO2 (e.g. in patients receiving noninvasive ventilation for hypercapnic respiratory failure). Finally, despite recommendations for systematic use during intubation in the ICU, the clinical impact of this strategy remains to date unknown in the ICU setting. Indeed, most of the recommendations are made from the NAP 4 audit (Cook et al. 2011), which showed that 74% of deaths related to airway issues in the ICU (tube displacement, oesophageal intubation) were associated with the lack of capnography.



Dead space measurement and volumetric capnography

Partial pressure of expired CO2 is usually plotted against time on a capnogram, allowing assessment of PEtCO2. Partial pressure of expired CO2 may also be plotted as a function of expired tidal volume to provide volumetric capnography. An ideal volumetric capnography curve can be described in three expiratory phases (Figure 1) (Verscheure et al. 2016): 


  • Phase 1 - exhaled CO2 amounts to zero and reflects the lack of CO2 content in the conducting airways
  • Phase 2 - CO2 increases linearly, reflecting mixing of CO2 content of distal airways and alveoli close to the main airways
  • Phase 3 - CO2 reaches a slowly rising plateau (reflecting alveolar gas compartment), whose slope is an indication of ventilation-perfusion mismatch. Both anatomical (VDaw) and alveolar dead space (VDalv) can be computed by combining capnography and arterial blood sampling using graphical analysis of the volumetric capnography curve (Fletcher et al. 1981). Assuming PaCO2 approximates PACO2, physiological dead space (VDphys) can also be computed using the Enghoff modification of the Bohr equation, as follows: 

            VDphys Bohr-Enghoff = (PaCO2-PACO2)/PaCOwith PACO2 = mean expired CO2 partial pressure 


However, VDphys assessed with the Bohr-Enghoff equation overestimates the true VDphys, since venous admixture and low ventilation-perfusion areas increase the difference between PaCO2 and PACO2. It was recently shown that the midportion of phase 3 is a reliable estimator of PACO2 (Tusman et al. 2011), allowing a continuous assessment of true physiological dead space, without arterial blood sampling, using Bohr’s original equation, as follows: 


VDphys Bohr = (PACO2-PeCO2)/PACO2 


VDphys Bohr as assessed by volumetric capnography has been shown to be closely related to dead space measurement using the multiple inert gas technique (Tusman et al. 2011), and is not impacted by the effect of shunt or low ventilation-perfusion areas.


Finally, plotting the volume of expired CO2 as a function of expired tidal volume allows computation of VDalv (Fletcher et al. 1981) and alveolar ejection volume (Romero et al. 1997), defined at the predicted point where alveolar emptying begins (Figure 2), as an attempt to better quantify phase 3 of the volumetric capnography curve.


The following clinical applications of volumetric capnography have been reported:

  • Nuckton et al. (2002) reported that dead space (computed from the Bohr-Enghoff equation) was strongly associated with acute respiratory distress syndrome (ARDS) mortality. In addition, Gattinoni et al. (2003) showed that decrease of dead space during prone position was associated with lower ARDS mortality. Alveolar ejection volume is also associated with ARDS mortality (Lucangelo et al. 2008), with the advantage of being independent of ventilatory settings (Romero et al. 1997). Whether strategies aiming to minimise deadspace decrease ARDS mortality remains however unknown.
  • As early as 1975, Suter et al. (1975) showed that the "best positive end-expiratory pressure (PEEP)" (i.e. PEEP level associated with the highest oxygen transport) was associated with the lowest dead space (computed from the Bohr-Enghoff equation with a correction for the effect of shunt [Kuwabara et al. 1969]). Increase in dead space below and above the best PEEP level was interpreted as evidence of lung overdistension (and hence compression of alveolar vessels) at low and high lung aerated volume.
  • Increased dead space fraction has been reported as an excellent predictor of extubation failure in a single study on ICU adult patients (González-Castro et al. 2011), a finding that should be confirmed before application in clinical practice.

However volumetric capnography presents the following limitations: measurement errors due to air leaks (e.g. during noninvasive ventilation) or obstruction of airway adaptor by secretions or condensation droplets, requirement of sufficient expiratory time in order to allow complete CO2 exhalation, deviation from the ideal 3 phases curve in some clinical situations, dependency to ventilatory settings etc.



PaCO2 surrogate

PEtCO2 can be used as a surrogate of PaCO2, e.g. after changing ventilatory settings, assuming the PEtCO2-PaCO2 gap remains constant over time (i.e. hypothesising that the dead space remains constant). It was shown as a cost-effective intervention (Rowan et al. 2015) with a reduction of a third of blood gases analyses and a saving of $880,496 over 6 months in an American paediatric ICU. In specific populations where tight control of PaCO2 is important (e.g. patients with brain injury), PEtCO2 monitoring can be useful as it allows a continuous noninvasive estimation of PaCO2. It should, however, be stressed that interventions that can alter dead space (e.g. bronchodilators, major change of PEEP level and so on) should prompt the verification of the PEtCO2-PaCO2 gap. Furthermore, it may be unsafe to use only PEtCO2 for the setting of the ventilator, given that some patients with unknown enlarged PEtCO2-PaCO2 gap might suffer from severe iatrogenic hypercapnia.

Capnography and haemodynamic intensive care


Cardiac arrest

The use of capnography in the cardiac arrest setting has been well documented and is recommended in both American and European resuscitation guidelines (Link et al. 2015; Soar et al. 2015). In this setting, capnography is a versatile tool that can help the management of cardiac arrest patients (Heradstveit et al. 2014).


First, as described above, it can confirm the correct placement of an endotracheal tube. Second, capnography may assess the quality of cardiac resuscitation, which is correlated with the PEtCO2 level, since PEtCO2 is related to cardiac output. Hence it is recommended (Paiva et al. 2018) to achieve a PEtCO2>20 mmHg during resuscitation. Third, capnography may help to detect the return to a spontaneous circulation if a sudden rise of PEtCO2 occurs. Fourth, initial low values or failure to maintain “correct” values of PEtCO2 are associated with worse outcome (Levine et al. 1997), and may help the decision to terminate resuscitation or help to triage patients with refractory cardiac arrest eligible for extracorporeal life support (Conseil français de réanimation cardiopulmonaire et al. 2009).


Cardiac output surrogate

As PEtCO2 is highly correlated to CO and monitored on a breath-by-breath basis (Weil et al. 1985), it may be used as a surrogate for continuous cardiac output monitoring over short periods, assuming CO2 production and elimination remain constant. In this connection, some authors recently investigated whether PEtCO2 variations could be used to track CO changes related to change in cardiac loading conditions (Monnet et al. 2012). In a study on 65 mechanically ventilated patients with acute circulatory failure, PEtCO2 increased by at least 5% during a passive leg raising manoeuvre predicted fluid responsiveness with 100% specificity and a sensitivity of 71%. For patients under veno-arterial extracorporeal membrane oxygenation (ECMO), PEtCO2 might reflect transpulmonary (or native) cardiac output. Naruke et al. (2010) reported that patients that were successfully weaned from veno-arterial ECMO exhibited a rise of PEtCO2 of at least 5 mmHg after reduction of ECMO flow, which was interpreted as a rise of native CO. If confirmed, this method could allow a more precise screening of patients who could be safely weaned from ECMO.


CO can be measured noninvasively with the differential Fick method, using measurements of CO2 elimination (VCO2) and PEtCO2 on a breath-by-breath basis before and during a CO2 rebreathing manoeuvre (Jaffe 1999). Since cardiac output is computed from expired CO2, blood flow from non-ventilated lung regions is not accounted for and a correction for shunt fraction has to be performed using arterial oxygen saturation assessed by a pulse oximeter and an iso-shunt diagram (Rocco et al. 2004). The NICO® system is a commercially available device based on this technique, using a rebreathing loop controlled by a pneumatic valve inserted at the Y-piece level. This technique has an acceptable reliability (Gueret et al. 2006) in cardiac surgery patients, but a lack of reliability in conditions of high intrapulmonary shunt (Rocco et al. 2004). In addition, the technique is hampered by several additional limitations: CO monitoring is not continuous (1 measurement every 3 minutes) making the technique unsuitable to assess fluid responsiveness by the passive leg raising test or other postural tests (Yonis et al. 2017); the technique is contraindicated in patients requiring strict control of PaCO2 (e.g. brain-injured patients); haemodynamic and respiratory instability (with rapidly changing VCO2 between the basal and rebreathing phase) may decrease the reliability of the CO measurement. The ideal patient for this technique would be mechanically ventilated, with no active breathing and no pulmonary disease, which probably makes the technique suitable for intraoperative monitoring.


Capnography for metabolic intensive care

As highlighted in the introduction, PEtCO2 is highly correlated to VCO2. The analysis of VCO2 and oxygen consumption (VO2) allows estimation of the energy expenditure, through indirect calorimetry (Oshima et al. 2017). This method is the gold standard for the estimation of the daily nutrition needs in the ICU, and is now implemented in some ICU ventilators. However, to be accurate, indirect calorimetry should be done under respiratory and haemodynamic stable conditions, in aerobic condition, and with FiO2 below 60%. Some authors have proposed a new estimation of energy expenditure using only VCO2 assessed by a commercial ventilator (Stapel et al. 2015), by computing respiratory quotient of the administered nutrition using the Weir formula to compute VO2 (Weir 1949). The reliability of this method was acceptable, with a less than 10% overestimation of energy expenditure. Whatever the method to determine energy expenditure, no study has, to date, explored the impact of its use on ICU outcome as compared to traditional predictive equations including anthropometric parameters.

Capnography for intrahospital transport

The use of PEtCO2 for intrahospital transport mechanically ventilated patients is highly recommended in the United Kingdom and in selected patients in France (Intensive Care Society 2016; Quenot et al. 2012). This recommendation is based on the frequency of airway related adverse events (hypoxia, extubation, ventilator failure etc.). As for the intubation procedure, strong evidence is lacking, and recommendations are based on observational studies and expert opinions.



In this review, we have highlighted the possible applications of capnography in the ICU. The versatility of this tool also makes its frailty. However, it often requires that the patient is haemodynamically stable, passively ventilated, with no change in energy expenditure. ICU patients do not fulfil these criteria most of the time. Finally, we are lacking studies showing improvement of critically ill patients’ outcome by using any of the CO2 monitoring tools. Whether this statement will remain true in the future depends on the will of critical care teams to embrace this technology, study it, and ultimately... use it!

Conflicts of interest

Mehdi Mezidi declares that he has no conflict of interest. Jean-Christophe Richard declares that he has no conflict of interest.


ARDS acute respiratory distress syndrome

CO cardiac output

CO2 carbon dioxide

ECMO extracorporeal membrane oxygenation

ICU intensive care unit

PACO2 CO2 alveolar partial pressure

PaCO2 CO2 arterial partial pressure

PeCO2 mean expired CO2 partial pressure 

PEEP positive end-expiratory pressure

PEtCO2 end-tidal CO2 partial pressure

VCO2 CO2 production

VDalv alveolar dead space

VDaw anatomical dead space

VDphys physiological dead space

VO2 O2 consumption


«« AKI: long-term impact on mortality and renal function

Vancomycin pharmacokinetics in critically ill obese patients »»


Conseil français de réanimation cardiopulmonaire et al. (2009) Guidelines for indications for the use of extracorporeal life support in refractory cardiac arrest. Annales Françaises d'Anesthésie et de Réanimation, 28(2): 187–90. 

Cook TM et al. (2011) Fourth National Audit Project. Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal 

College of Anaesthetists and the Difficult Airway Society. Part 2: intensive care and emergency departments. Br J Anaesth, 106(5): 632–42. 

Dunn SM et al. (1990) Tracheal intubation is not invariably confirmed by capnography. Anesthesiology. 73(6): 1285–7. 

Fletcher R et al. (1981) The concept of deadspace with special reference to the single breath test for carbon dioxide. Br J Anaesth, 53(1):77–88. 

Gattinoni L et al. (2003) Decrease in PaCO2 with prone position is predictive of improved outcome in acute respiratory distress syndrome. Crit. Care Med, 31(12): 2727–33. 

Georgiou AP et al. (2010) The use of capnography and the availability of airway equipment on Intensive Care Units in the UK and the Republic of Ireland. Anaesthesia, 65(5): 462–7. 

González-Castro A et al. (2011) [Utility of the dead space fraction (Vd/Vt) as a predictor of extubation success]. Med Intensiva, 35(9): 529–38. 

Gueret G et al. (2006) Cardiac output measurements in off-pump coronary surgery. Eur J of Anaesthesiol. 23(10): 848–54. 

Guggenberger H et al. (1989) Early detection of inadvertent oesophageal intubation: pulse oximetry vs. capnography. Acta Anaesth Scand, 33(2): 112–5. 

Heradstveit BE et al. (2012) Factors complicating interpretation of capnography during advanced life support in cardiac arrest—A clinical retrospective study in 575 patients. Resuscitation. 83(7): 813–8. 

Heradstveit BE et al. (2014) PQRST – A unique aide-memoire for capnography interpretation during cardiac arrest. Resuscitation, 85(11): 1619–20. 

Intensive Care Society (2016) 2016 Update of the Intensive Care Society standards for Capnography in Critical Care. [Internet]. 2017 [Accessed: 15 May 2018]. pp. 1–6. Available from:

Jaffe MB (1999) Partial CO2 rebreathing cardiac output--operating principles of the NICO system. J Clin Monit Comput, 15(6): 387–401. 

Kuwabara S et al. (1969). Effect of anatomic shunt on physiologic deadspace-to-tidal volume ratio--a new equation. Anesthesiology, 31(6): 575–7. 

Levine RL et al. (1997) End-tidal carbon dioxide and outcome of out-of-hospital cardiac arrest. N Engl J Med, 337(5): 301–6. 

Link MS et al. (2015) Part 7: Adult Advanced Cardiovascular Life Support: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation,132 (18 Suppl 2): S444–64. 

Lucangelo U et al. (2008) Prognostic value of different dead space indices in mechanically ventilated patients with acute lung injury and ARDS. Chest, 133(1): 62–71. 

Monnet X et al. (2012) Passive leg-raising and end-expiratory occlusion tests perform better than pulse pressure variation in patients with low respiratory system compliance. Crit. Care Med, 40(1): 152–7. 

Naruke T et al. (2010) End-tidal carbon dioxide concentration can estimate the appropriate timing for weaning off from extracorporeal membrane oxygenation for refractory circulatory failure. Int Heart J, 51(2):116–20. 

Nuckton TJ et al (2002) Pulmonary dead-space fraction as a risk factor for death in the acute respiratory distress syndrome. N Engl J Med, 346(17): 1281–6. 

Ono Y et al. (2016) Difficult airway management resources and capnography use in Japanese intensive care units: a nationwide cross-sectional study. J Anesth, 30(4): 644–52. 

Oshima T et al. (2017) Indirect calorimetry in nutritional therapy. A position paper by the ICALIC study group. Clin Nutr, 36(3): 651–62. 
Paiva EF et al. (2018) The use of end-tidal carbon dioxide (ETCO2) measurement to guide management of cardiac arrest: A systematic review. Resuscitation, 123:1–7. 

Quenot J-P et al. (2012) Intrahospital transport of critically ill patients (excluding newborns) Recommendations of the Societe de Reanimation de Langue Francaise (SRLF), the Societe Francaise d“Anesthesie et de Reanimation (SFAR), and the Societe Francaise de Medecine d”Urgence (SFMU). Ann Intensive Care, 2(1):1. 

Rocco M et al. (2004) A comparative evaluation of thermodilution and partial CO2 rebreathing techniques for cardiac output assessment in critically ill patients during assisted ventilation. Intensive Care Med, 30(1): 82–7. 

Romero PV et al. (1997) Physiologically based indices of volumetric capnography in patients receiving mechanical ventilation. Eur. Respir. J, 10(6): 1309–15. 

Rowan CM et al. (2015) Implementation of Continuous Capnography Is Associated With a Decreased Utilization of Blood Gases. J Clin Med Res, 7(2): 71–5. 

Silvestri S et al. (2005) The effectiveness of out-of-hospital use of continuous end-tidal carbon dioxide monitoring on the rate of unrecognized misplaced intubation within a regional emergency medical services system. Ann Emerg Med, 45(5):497–503. 

Soar J et al. (2015) European Resuscitation Council Guidelines for Resuscitation 2015: Section 3. Adult advanced life support. Resuscitation, 95:100-47 

Stapel SN et al. (2015) Ventilator-derived carbon dioxide production to assess energy expenditure in critically ill patients: proof of concept. Crit Care, 19:370. 

Suter PM et al. (1975) Optimum end-expiratory airway pressure in patients with acute pulmonary failure. N Engl J Med, 292(6): 284–9. 

Tusman G et al. (2011) Validation of Bohr dead space measured by volumetric capnography. Intensive Care Med, 37(5): 870–4. 

Verscheure S et al. (2016) Volumetric capnography: lessons from the past and current clinical applications. Crit Care, 20(1): 184. 

Weil MH et al. (1985) Cardiac output and end-tidal carbon dioxide. Crit. Care Med,13(11): 907–9. 

Weir JB de V. (1949) New methods for calculating metabolic rate with special reference to protein metabolism. J Physiol, 109(1-2):1–9.

Yonis H et al. (2017) Change in cardiac output during Trendelenburg maneuver is a reliable predictor of fluid responsiveness in patients with acute respiratory distress syndrome in the prone position under protective ventilation. Crit Care, 21(1): 295. 

Related Articles

Advanced haemodynamic monitoring methods   Bernd Saugel, MD, EDIC is a Professor of Anesthesiology and works as a consultant... Read more

When envisioning the future of haemodynamic monitoring, we cannot limit the discussion to new sensors and computer innovations.... Read more

Critical Care Medicine has existed for many years, but was only recognised as a specialty in the last 40 years or so. However,... Read more

Related IssueArticles

In this article we focus on the evidence of whole-body ultrasonography used for hypotension or shock. We first highlight individual... Read more

A critical care pharmacist’s perspective and advice on medication safety around sedative and analgesic therapy in the... Read more

A review of diagnosis and treatment of dengue, a mosquito-borne febrile illness caused by flavivirus with a clinical spectrum... Read more

capnography, Haemodynamic monitoring, Respiratory monitoring Reviews the potential uses and pitfalls of capnography in critically ill patients, especially for haemodynamic and respiratory monitoring. Expired CO2 c...

No comment

Please login to leave a comment...

Highlighted Products