Accepted Manuscript Capnography During Critical Illness Boulos S. Nassar, MD, Gregory A. Schmidt, MD, FCCP PII:
S0012-3692(15)00048-3
DOI:
10.1378/chest.15-1369
Reference:
CHEST 47
To appear in:
CHEST
Received Date: 5 June 2015 Revised Date:
15 September 2015
Accepted Date: 16 September 2015
Please cite this article as: Nassar BS, Schmidt GA, Capnography During Critical Illness, CHEST (2015), doi: 10.1378/chest.15-1369. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
RI PT
ACCEPTED MANUSCRIPT
Capnography During Critical Illness
And
SC
Boulos S. Nassar MD
TE D
M AN U
Gregory A. Schmidt, MD, FCCP
AC C
EP
Corresponding Address: Gregory A. Schmidt, MD Division of Pulmonary Diseases, Critical Care, and Occupational Medicine Department of Internal Medicine: C33-GH University of Iowa 200 Hawkins Drive Iowa City, IA 52242 [email protected]
The authors have no conflicts to disclose with regards to the content of this manuscript.
ACCEPTED MANUSCRIPT
Abstract
AC C
EP
TE D
M AN U
SC
RI PT
Capnography has made steady inroads in the ICU, being increasingly used for all mechanically ventilated patients. There is growing recognition that capnography is rich in information about lung and circulatory physiology, providing insight into many diseases and treatments. These include conditions of impaired matching of ventilation and perfusion, such as pulmonary embolism and obstructive lung diseases; circulatory questions such as the adequacy of chest compressions during cardiac arrest or fluidresponsiveness in patients in shock; and the safety of procedural sedation. In this review, we emphasize analysis of the entire capnographic waveform as a way to glean additional useful information. At the same time, we discuss important limitations of capnography, especially when it is considered as a surrogate for the PaCO2.
ACCEPTED MANUSCRIPT
Volume- and Time-Capnography
M AN U
SC
RI PT
Capnography refers to the measurement and display of carbon dioxide (CO2) concentrations in respiratory gases. By the 1990’s capnography was widely available for critically ill patients and quickly became part of routine monitoring. There is broad appreciation of its value for assuring that endotracheal tubes are placed properly in the trachea1; it is also being used to guide tidal volume and rate settings during mechanical ventilation, focusing on the end-tidal value. Yet the full capnogram is much richer in information than generally appreciated, finding application to gauge the degree of ventilation (V)-perfusion (Q) mismatch; measure dead space and quantify airflow obstruction in asthma and chronic obstructive pulmonary disease (COPD); diagnose pulmonary embolism (PE) and distinguish it from exacerbations of chronic obstructive pulmonary disease; judge the adequacy of chest compressions in cardiac arrest and detect return of spontaneous circulation (ROSC); estimate changes in cardiac output; predict fluid responsiveness; and assist in metabolic assessment and nutritional needs2. A full appreciation of capnography is founded on the physiology of CO2 exchange in the lung. Accordingly, this review defines volume-capnography versus time-capnography; describes the phases of the capnogram; explores the determinants of the CO2 concentration in a single alveolus and in expired gas; addresses the correlation between arterial and end-tidal CO2 (ET-CO2) values in ventilated patients; and reviews several key clinical applications relevant to the critically ill.
AC C
EP
TE D
Capnography refers to the measurement of CO2 at the airway opening, generally displayed as a partial pressure (PCO2). Colorimetric, qualitative capnometers can be placed directly in the airway to confirm endotracheal tube placement, but will not be discussed here further.3,4 Quantitative capnography in the intensive care unit (ICU) typically relies on infrared absorption, placing the sensor directly in the ventilator circuit (mainstream capnography) or continuously drawing a sample from the airway and directing it through fine-bore tubing to an off-line sensor (sidestream capnography). Sidestream capnography can also be sampled through a nasal cannula in spontaneously breathing patients, while mainstream capnography is acquired only from an endotracheal or tracheostomy tube. The measured PCO2 can be displayed as a function of time or of expired volume (time- and volume-capnography, respectively, Figs 1 A and B). Because it relates CO2 to expired volume, volume-capnography more intuitively links anatomy to PCO2 and allows calculation of dead space and CO2 production, but requires measuring flow instantaneously5. This entails higher cost, a mainstream sensor, and more complexity, precluding its wide application for routine ICU monitoring currently, although it is available in some commercial ventilators and as standalone monitors. On the other hand, time capnography is readily available in most ICUs but does not take expiratory flows into account. In this brief review we focus on standard ICU capnography displaying the time-capnogram, as shown in Figure 1A. Phases of the Time-Capnogram The capnogram is divided into four phases (fig 1A). In phase I the PCO2 is effectively zero, representing the portion of inspiration during which fresh, CO2-free gas enters the airway, and early expiration while dead space gas (also CO2-free) is exhaled. PCO2 should always be zero unless there is rebreathing of CO2laden expired gas6. Phase II is more interesting, mixing the vestiges of anatomic dead space with alveolar gas, coming from both alveolar dead space and ventilated alveoli. In addition, phase II includes some gas
ACCEPTED MANUSCRIPT
that truly “belongs” to phases I and III due to turbulent mixing and diffusion at the interface of dead space and alveolar gas. For these reasons, phase II rises abruptly when V and Q are well-matched, but more gradually when some high V/Q alveoli (having low PCO2) contribute to the expired gas stream. Thus many diffuse lung and pulmonary vascular diseases are associated with slowly rising phase II.
M AN U
SC
RI PT
Phase III represents the alveolar plateau, displaying PCO2 values largely from alveoli. In a single alveolus at steady state, CO2 concentration represents the balance between ventilation, which tends to lower it, and perfusion, which tends to raise it. In the ideal lung (where ventilation and perfusion are matched and all alveoli are identical), mean alveolar partial pressure for CO2 (PACO2) is equal to the arterial partial pressure for CO2 (PaCO2). Breathing is phasic, however, so that the instantaneous PACO2 varies, falling as fresh gas enters the alveolar compartment and rising again throughout expiration7,8 (Figure 2). The maximal PACO2 approaches the venous (not arterial) PCO2 level, a value roughly 7 mmHg higher than arterial. In theory, steady-state ET-CO2 can range from as high as the venous PCO2 (when the lung is healthy and expiratory time is sufficiently long) to very, very low values (when overall ventilation is high in comparison to VCO2; in the presence of alveolar dead space; and when expiratory time is short). These are some of the many reasons that ET-CO2 may correlate very poorly with PaCO2 (Figure 3; discussed further below), especially in the critically ill.
TE D
Although the normal phase III appears relatively flat, the PCO2 rises progressively, in part reflecting the fact that alveolar PCO2 rises during expiration as pulmonary blood flow continues to offload CO2 during expiration into an ever-decreasing alveolar volume, but also for many other reasons in both health and disease (Table 1). Phase III also rises more quickly when CO2 production is high, as shown elegantly by DuBois and colleagues who measured their own capnograms during exercise nearly 70 years ago.7 Cardiac output only perturbs the phase III rise or affects the end-tidal value when changing, as following a fluid bolus or when spontaneous circulation is restored during resuscitation. The most clinically important contributor to the phase III slope is V/Q heterogeneity related to lung disease. Finally, phase IV, like phase I, is trivial, plunging rapidly to zero shortly after the onset of inspiration.
AC C
EP
The volume capnogram presents CO2 concentration in relation to expired volume and is divided into 3 similar phases (lacking a phase IV since inspiration is not plotted; figure 1B). Phase I refers to the emptying of the anatomic dead space. Phase II reflects the transition from airway dead space to proximal alveoli, while phase III refers to alveolar emptying. The significance of the shapes of these three phases is similar to those of time-capnography, but the slope of phase III is steeper because expiratory flow falls exponentially throughout expiration.
Clinical Applications
Correlation between ET-CO2 and PaCO2 during mechanical ventilation For adjusting mechanical ventilation settings, time capnography offers an attractive alternative to arterial blood gas analysis, being continuous and non-invasive. Yet while ET-CO2 correlates well with PaCO2 in health, large discrepancies are common in disease, often related to the factors shown in Table
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
1.9,10,11,12 ET-CO2 values may exceed PaCO2 or be very much lower13. When ventilator settings are adjusted, even the changes in ET-CO2 and PaCO2 are poorly related, even moving in opposing directions14 (Figure 3). Further, trends over time show a poor correlation between changes in ET-CO2 and PaCO2.15,16 Perhaps this should not be surprising when all the determinants of ET-CO2 are considered. For example, augmenting ventilator rate or tidal volume should tend to lower both ET-CO2 (because it tends to raise the ratio of alveolar ventilation to CO2 production) and PaCO2 but, if dead space fraction rises, ET-CO2 may fall while PaCO2 rises. Because PEEP affects alveolar dead space, its value impacts the capnogram: moreover, these effects have been used to guide PEEP levels in ARDS. 17 18 19 20 These data show that relying on ET-CO2 as a surrogate for PaCO2 is prone to serious error. Nevertheless, while ventilator waveforms offer similar monitoring, capnography remains important during mechanical ventilation to ensure airway patency and detect dislodgement or disconnection of the endotracheal tube.21 It probably plays a bigger role during patient transport when the risk of endotracheal tube migration is even higher.
Pulmonary embolus
TE D
Pulmonary embolism (PE) remains a diagnostic challenge in the ICU. While computed tomographic angiography (CTA) is reasonably sensitive and specific, critically ill patients present difficulties in safe transport, often have many alternative reasons for new cardiopulmonary events, and are at risk of acute kidney injury, making CTA a cumbersome screening test. PE compromises perfusion of the affected alveoli, with much less impact on their ventilation, suggesting that capnography could be diagnostically useful. Based on the physiology described above, one would expect PE to a) raise the alveolar dead space (thus the physiologic and alveolar dead space fractions); b) lower the expired CO2 value throughout phase III and widen any gradient between PaCO2 and ET-CO2 (since gas expired from high V/Q alveoli has little CO2); and c) not raise the phase III slope (perhaps allowing distinction from COPD and other airway diseases with marked V/Q heterogeneity; Figure 4).
AC C
EP
Volume capnography, as opposed to time capnography, links expired CO2 values more closely to lung volumes and allows calculation of physiologic dead space fractions (by using the modified Bohr equation and a simultaneous arterial blood gas). In testing the diagnostic value in patients with suspected PE, most studies have employed volume capnography, but similar principles underlie the changes seen with time capnography.22 In one of the earliest studies, PE was shown to raise physiologic dead space fraction and the arterial-end-tidal CO2 difference.23 These findings have been confirmed in subsequent studies.24 25 In addition, the alveolar dead-space fraction correlates with the lung perfusion defect on lung scintigraphy and the pulmonary artery pressure at angiography.25 Any difference between arterial and ET-CO2 depends on tidal volume, leading several investigators to standardize the point of measurement by extrapolating phase III to a tidal volume equal to 15% of predicted total lung capacity. The difference between the expired CO2 value and PaCO2 at this extrapolated volume lends itself to calculation of the “late dead-space fraction”. When it exceeds 12-15%, this value, even more than overall dead-space fraction or ET-CO2 to PaCO2 gradient, may best separate those with and without PE. 26 27 28 29 30 Others
ACCEPTED MANUSCRIPT
have combined the late dead-space fraction or ET-CO2:PaCO2 gradient with a negative D-dimer value to reduce the post-test probability of PE sufficiently to obviate further diagnostic testing.25
RI PT
Differences in PaCO2 and ET-CO2 are common in critical illness yet only occasionally is this caused by PE. Most alternative causes consist of lung diseases with great V/Q heterogeneity, having abnormally steep rises during phase III for the reasons cited above. In contrast, PE should not steepen phase III (in fact, it flattens the slope for complex reasons related to redistribution of blood flow) 31; and thrombolysis has been shown to increase the slope.32 In a series of patients presenting to the emergency department with suspected PE, the slope of phase III was significantly flatter in those with confirmed PE.26
Obstructive Lung Diseases:
TE D
M AN U
SC
In a pooled analysis of 14 trials and 2291 subjects, the sensitivity of capnography for the diagnosis of PE was 0.80 while the specificity was 0.49.33 The diagnostic use of capnography for suspected PE has several limitations. First, most clinical data pertain to volume capnography: time capnography does not allow calculation of the late dead-space fraction at a standardized tidal volume. A second issue regards steady-state -- PE may perturb the circulation, causing changes in CO2 delivery to the lung that invalidate the assumption of steady-state. Moreover, most studies have been conducted on clinically stable patients in the emergency department, rather than those mechanically ventilated in an ICU. Finally, since changes in ventilator settings or patient-ventilator interaction can change the capnographic waveform, care must be taken that artifacts of care are not confused with signals of pathology. Nevertheless, in the right clinical setting, the capnogram may provide clues to raise or lower suspicion for PE, especially if joined to a negative D-dimer value.
AC C
EP
Obstructive lung diseases are characterized by widely varying V/Q ratios, producing typical capnographic signatures (Fig 4).34 The phase II transition from anatomic dead space to alveolar gas is blurred by the early contribution from high V/Q units (having low PACO2), tending to reduce the steepness of its rise. During phase III, units with varying V/Q ratios continue to empty in desynchronized fashion (with higher V/Q regions - having lower PACO2 - contributing more earlier, while lower V/Q regions having higher PACO2 dominating later), amplifying the rise of the plateau phase. This combination causes the alpha angle between phases II and III to increase35. In severe obstructive diseases, the capnogram assumes a “shark’s fin” appearance (Figure 5a). The steeper phase III also causes the end-tidal value to depend more strongly on expiratory time. This feature, combined with the anatomic dead space inherent in emphysema, causes marked discrepancies between end-tidal and arterial PCO2 values. These qualitative features of obstructive diseases on time-capnography correspond to quantitative aspects on volume-capnography. An increased phase III slope has been demonstrated in most diseases with airflow obstruction (COPD, asthma and bronchiectasis) 36, 37. The degree of slope correlates with the severity of airflow obstruction measured by spirometry36,37,38,39,40,41 and the degree of emphysema seen on chest CT 42. Automated analysis of multiple capnographic indices may increase the sensitivity and specificity for diagnosing obstructive lung disease. Combining exhalation duration, ET-CO2, endexhalation slope, and time spent at ET-CO2, succeeds in distinguishing patients with COPD from those
ACCEPTED MANUSCRIPT
with congestive heart failure.43 Capnography may be even more sensitive than spirometry for detecting early, small airway changes, as has also been demonstrated for the similar technique of single-breath nitrogen washout.44,45
SC
RI PT
ET-CO2 may differ markedly from the simultaneous PaCO2, especially in obstructed patients, and this is true when patients are stable or having an exacerbation.46,47 A slow, forced, maximal expiration can produce a much better correlation between ET-CO2 and PaCO246, but this is not practical in the critically ill. In some ventilated patients with COPD the difference between ET-CO2 and PaCO2 can exceed 40mmHg, emphasizing the serious errors that can be made in drawing inferences about the adequacy of ventilation based on ET-CO2 values.
Capnography and cardiac output:
It is commonly recognized that ET-CO2 values change with the return of spontaneous circulation (ROSC) ; when passive leg raising (PLR) augments cardiac output51, 52, 53; and in many shock states54, 55. These findings give rise to the incorrect idea that cardiac output directly influences ET-CO2. In steady state, however, the volume of CO2 exhaled through the lungs must equal that produced in the tissues metabolically, regardless of cardiac output56. Cardiac output largely affects ET-CO2 only in dynamic situations, not when the circulation is stable.
M AN U
48,49, 50
EP
TE D
The appearance of a direct relationship between cardiac output and ET-CO2 relates to non-steady state conditions combined with the fact that the body can store large amounts of CO2.55 57 58For example, if cardiac output is reduced through hemorrhagic shock, ET-CO2 falls transiently as less CO2 returns to the lungs. At the lower cardiac output, tissue and venous PCO2 rise, ultimately restoring the total volume of CO2 reaching the lungs, balancing metabolic production with excretion. Because so much CO2 can be stored in the body as tissue CO2 rises, equilibrium will not be reached quickly. In addition, many interventions to change cardiac output also change dead space (for example, by raising or lowering left atrial pressure).59 60 Thus short-term experiments may give the appearance of a direct relationship of cardiac output and time-capnography.
AC C
There is a more subtle effect of cardiac output on alveolar and ET-CO2 related to the Fick Principle. Recalling that alveolar CO2 tends to approximate the venous value at end-expiration, it should be clear that low cardiac output states will produce larger (but still modest) swings in alveolar and ET-CO 2. The tendency, then, is for low cardiac output to produce higher (not lower) ET-CO2 in steady-state and when all other variables are kept constant. Although ET-CO2 does not directly reflect cardiac output, by signaling changes, it has great value during resuscitation and when judging fluid-responsiveness. Cardiopulmonary resuscitation (CPR): During CPR of cardiac arrest, capnography signifies increasing pulmonary blood flow, thus the adequacy of chest compressions.61 62 An ET-CO2 value of
S0012-3692(15)00048-3
DOI:
10.1378/chest.15-1369
Reference:
CHEST 47
To appear in:
CHEST
Received Date: 5 June 2015 Revised Date:
15 September 2015
Accepted Date: 16 September 2015
Please cite this article as: Nassar BS, Schmidt GA, Capnography During Critical Illness, CHEST (2015), doi: 10.1378/chest.15-1369. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
RI PT
ACCEPTED MANUSCRIPT
Capnography During Critical Illness
And
SC
Boulos S. Nassar MD
TE D
M AN U
Gregory A. Schmidt, MD, FCCP
AC C
EP
Corresponding Address: Gregory A. Schmidt, MD Division of Pulmonary Diseases, Critical Care, and Occupational Medicine Department of Internal Medicine: C33-GH University of Iowa 200 Hawkins Drive Iowa City, IA 52242 [email protected]
The authors have no conflicts to disclose with regards to the content of this manuscript.
ACCEPTED MANUSCRIPT
Abstract
AC C
EP
TE D
M AN U
SC
RI PT
Capnography has made steady inroads in the ICU, being increasingly used for all mechanically ventilated patients. There is growing recognition that capnography is rich in information about lung and circulatory physiology, providing insight into many diseases and treatments. These include conditions of impaired matching of ventilation and perfusion, such as pulmonary embolism and obstructive lung diseases; circulatory questions such as the adequacy of chest compressions during cardiac arrest or fluidresponsiveness in patients in shock; and the safety of procedural sedation. In this review, we emphasize analysis of the entire capnographic waveform as a way to glean additional useful information. At the same time, we discuss important limitations of capnography, especially when it is considered as a surrogate for the PaCO2.
ACCEPTED MANUSCRIPT
Volume- and Time-Capnography
M AN U
SC
RI PT
Capnography refers to the measurement and display of carbon dioxide (CO2) concentrations in respiratory gases. By the 1990’s capnography was widely available for critically ill patients and quickly became part of routine monitoring. There is broad appreciation of its value for assuring that endotracheal tubes are placed properly in the trachea1; it is also being used to guide tidal volume and rate settings during mechanical ventilation, focusing on the end-tidal value. Yet the full capnogram is much richer in information than generally appreciated, finding application to gauge the degree of ventilation (V)-perfusion (Q) mismatch; measure dead space and quantify airflow obstruction in asthma and chronic obstructive pulmonary disease (COPD); diagnose pulmonary embolism (PE) and distinguish it from exacerbations of chronic obstructive pulmonary disease; judge the adequacy of chest compressions in cardiac arrest and detect return of spontaneous circulation (ROSC); estimate changes in cardiac output; predict fluid responsiveness; and assist in metabolic assessment and nutritional needs2. A full appreciation of capnography is founded on the physiology of CO2 exchange in the lung. Accordingly, this review defines volume-capnography versus time-capnography; describes the phases of the capnogram; explores the determinants of the CO2 concentration in a single alveolus and in expired gas; addresses the correlation between arterial and end-tidal CO2 (ET-CO2) values in ventilated patients; and reviews several key clinical applications relevant to the critically ill.
AC C
EP
TE D
Capnography refers to the measurement of CO2 at the airway opening, generally displayed as a partial pressure (PCO2). Colorimetric, qualitative capnometers can be placed directly in the airway to confirm endotracheal tube placement, but will not be discussed here further.3,4 Quantitative capnography in the intensive care unit (ICU) typically relies on infrared absorption, placing the sensor directly in the ventilator circuit (mainstream capnography) or continuously drawing a sample from the airway and directing it through fine-bore tubing to an off-line sensor (sidestream capnography). Sidestream capnography can also be sampled through a nasal cannula in spontaneously breathing patients, while mainstream capnography is acquired only from an endotracheal or tracheostomy tube. The measured PCO2 can be displayed as a function of time or of expired volume (time- and volume-capnography, respectively, Figs 1 A and B). Because it relates CO2 to expired volume, volume-capnography more intuitively links anatomy to PCO2 and allows calculation of dead space and CO2 production, but requires measuring flow instantaneously5. This entails higher cost, a mainstream sensor, and more complexity, precluding its wide application for routine ICU monitoring currently, although it is available in some commercial ventilators and as standalone monitors. On the other hand, time capnography is readily available in most ICUs but does not take expiratory flows into account. In this brief review we focus on standard ICU capnography displaying the time-capnogram, as shown in Figure 1A. Phases of the Time-Capnogram The capnogram is divided into four phases (fig 1A). In phase I the PCO2 is effectively zero, representing the portion of inspiration during which fresh, CO2-free gas enters the airway, and early expiration while dead space gas (also CO2-free) is exhaled. PCO2 should always be zero unless there is rebreathing of CO2laden expired gas6. Phase II is more interesting, mixing the vestiges of anatomic dead space with alveolar gas, coming from both alveolar dead space and ventilated alveoli. In addition, phase II includes some gas
ACCEPTED MANUSCRIPT
that truly “belongs” to phases I and III due to turbulent mixing and diffusion at the interface of dead space and alveolar gas. For these reasons, phase II rises abruptly when V and Q are well-matched, but more gradually when some high V/Q alveoli (having low PCO2) contribute to the expired gas stream. Thus many diffuse lung and pulmonary vascular diseases are associated with slowly rising phase II.
M AN U
SC
RI PT
Phase III represents the alveolar plateau, displaying PCO2 values largely from alveoli. In a single alveolus at steady state, CO2 concentration represents the balance between ventilation, which tends to lower it, and perfusion, which tends to raise it. In the ideal lung (where ventilation and perfusion are matched and all alveoli are identical), mean alveolar partial pressure for CO2 (PACO2) is equal to the arterial partial pressure for CO2 (PaCO2). Breathing is phasic, however, so that the instantaneous PACO2 varies, falling as fresh gas enters the alveolar compartment and rising again throughout expiration7,8 (Figure 2). The maximal PACO2 approaches the venous (not arterial) PCO2 level, a value roughly 7 mmHg higher than arterial. In theory, steady-state ET-CO2 can range from as high as the venous PCO2 (when the lung is healthy and expiratory time is sufficiently long) to very, very low values (when overall ventilation is high in comparison to VCO2; in the presence of alveolar dead space; and when expiratory time is short). These are some of the many reasons that ET-CO2 may correlate very poorly with PaCO2 (Figure 3; discussed further below), especially in the critically ill.
TE D
Although the normal phase III appears relatively flat, the PCO2 rises progressively, in part reflecting the fact that alveolar PCO2 rises during expiration as pulmonary blood flow continues to offload CO2 during expiration into an ever-decreasing alveolar volume, but also for many other reasons in both health and disease (Table 1). Phase III also rises more quickly when CO2 production is high, as shown elegantly by DuBois and colleagues who measured their own capnograms during exercise nearly 70 years ago.7 Cardiac output only perturbs the phase III rise or affects the end-tidal value when changing, as following a fluid bolus or when spontaneous circulation is restored during resuscitation. The most clinically important contributor to the phase III slope is V/Q heterogeneity related to lung disease. Finally, phase IV, like phase I, is trivial, plunging rapidly to zero shortly after the onset of inspiration.
AC C
EP
The volume capnogram presents CO2 concentration in relation to expired volume and is divided into 3 similar phases (lacking a phase IV since inspiration is not plotted; figure 1B). Phase I refers to the emptying of the anatomic dead space. Phase II reflects the transition from airway dead space to proximal alveoli, while phase III refers to alveolar emptying. The significance of the shapes of these three phases is similar to those of time-capnography, but the slope of phase III is steeper because expiratory flow falls exponentially throughout expiration.
Clinical Applications
Correlation between ET-CO2 and PaCO2 during mechanical ventilation For adjusting mechanical ventilation settings, time capnography offers an attractive alternative to arterial blood gas analysis, being continuous and non-invasive. Yet while ET-CO2 correlates well with PaCO2 in health, large discrepancies are common in disease, often related to the factors shown in Table
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
1.9,10,11,12 ET-CO2 values may exceed PaCO2 or be very much lower13. When ventilator settings are adjusted, even the changes in ET-CO2 and PaCO2 are poorly related, even moving in opposing directions14 (Figure 3). Further, trends over time show a poor correlation between changes in ET-CO2 and PaCO2.15,16 Perhaps this should not be surprising when all the determinants of ET-CO2 are considered. For example, augmenting ventilator rate or tidal volume should tend to lower both ET-CO2 (because it tends to raise the ratio of alveolar ventilation to CO2 production) and PaCO2 but, if dead space fraction rises, ET-CO2 may fall while PaCO2 rises. Because PEEP affects alveolar dead space, its value impacts the capnogram: moreover, these effects have been used to guide PEEP levels in ARDS. 17 18 19 20 These data show that relying on ET-CO2 as a surrogate for PaCO2 is prone to serious error. Nevertheless, while ventilator waveforms offer similar monitoring, capnography remains important during mechanical ventilation to ensure airway patency and detect dislodgement or disconnection of the endotracheal tube.21 It probably plays a bigger role during patient transport when the risk of endotracheal tube migration is even higher.
Pulmonary embolus
TE D
Pulmonary embolism (PE) remains a diagnostic challenge in the ICU. While computed tomographic angiography (CTA) is reasonably sensitive and specific, critically ill patients present difficulties in safe transport, often have many alternative reasons for new cardiopulmonary events, and are at risk of acute kidney injury, making CTA a cumbersome screening test. PE compromises perfusion of the affected alveoli, with much less impact on their ventilation, suggesting that capnography could be diagnostically useful. Based on the physiology described above, one would expect PE to a) raise the alveolar dead space (thus the physiologic and alveolar dead space fractions); b) lower the expired CO2 value throughout phase III and widen any gradient between PaCO2 and ET-CO2 (since gas expired from high V/Q alveoli has little CO2); and c) not raise the phase III slope (perhaps allowing distinction from COPD and other airway diseases with marked V/Q heterogeneity; Figure 4).
AC C
EP
Volume capnography, as opposed to time capnography, links expired CO2 values more closely to lung volumes and allows calculation of physiologic dead space fractions (by using the modified Bohr equation and a simultaneous arterial blood gas). In testing the diagnostic value in patients with suspected PE, most studies have employed volume capnography, but similar principles underlie the changes seen with time capnography.22 In one of the earliest studies, PE was shown to raise physiologic dead space fraction and the arterial-end-tidal CO2 difference.23 These findings have been confirmed in subsequent studies.24 25 In addition, the alveolar dead-space fraction correlates with the lung perfusion defect on lung scintigraphy and the pulmonary artery pressure at angiography.25 Any difference between arterial and ET-CO2 depends on tidal volume, leading several investigators to standardize the point of measurement by extrapolating phase III to a tidal volume equal to 15% of predicted total lung capacity. The difference between the expired CO2 value and PaCO2 at this extrapolated volume lends itself to calculation of the “late dead-space fraction”. When it exceeds 12-15%, this value, even more than overall dead-space fraction or ET-CO2 to PaCO2 gradient, may best separate those with and without PE. 26 27 28 29 30 Others
ACCEPTED MANUSCRIPT
have combined the late dead-space fraction or ET-CO2:PaCO2 gradient with a negative D-dimer value to reduce the post-test probability of PE sufficiently to obviate further diagnostic testing.25
RI PT
Differences in PaCO2 and ET-CO2 are common in critical illness yet only occasionally is this caused by PE. Most alternative causes consist of lung diseases with great V/Q heterogeneity, having abnormally steep rises during phase III for the reasons cited above. In contrast, PE should not steepen phase III (in fact, it flattens the slope for complex reasons related to redistribution of blood flow) 31; and thrombolysis has been shown to increase the slope.32 In a series of patients presenting to the emergency department with suspected PE, the slope of phase III was significantly flatter in those with confirmed PE.26
Obstructive Lung Diseases:
TE D
M AN U
SC
In a pooled analysis of 14 trials and 2291 subjects, the sensitivity of capnography for the diagnosis of PE was 0.80 while the specificity was 0.49.33 The diagnostic use of capnography for suspected PE has several limitations. First, most clinical data pertain to volume capnography: time capnography does not allow calculation of the late dead-space fraction at a standardized tidal volume. A second issue regards steady-state -- PE may perturb the circulation, causing changes in CO2 delivery to the lung that invalidate the assumption of steady-state. Moreover, most studies have been conducted on clinically stable patients in the emergency department, rather than those mechanically ventilated in an ICU. Finally, since changes in ventilator settings or patient-ventilator interaction can change the capnographic waveform, care must be taken that artifacts of care are not confused with signals of pathology. Nevertheless, in the right clinical setting, the capnogram may provide clues to raise or lower suspicion for PE, especially if joined to a negative D-dimer value.
AC C
EP
Obstructive lung diseases are characterized by widely varying V/Q ratios, producing typical capnographic signatures (Fig 4).34 The phase II transition from anatomic dead space to alveolar gas is blurred by the early contribution from high V/Q units (having low PACO2), tending to reduce the steepness of its rise. During phase III, units with varying V/Q ratios continue to empty in desynchronized fashion (with higher V/Q regions - having lower PACO2 - contributing more earlier, while lower V/Q regions having higher PACO2 dominating later), amplifying the rise of the plateau phase. This combination causes the alpha angle between phases II and III to increase35. In severe obstructive diseases, the capnogram assumes a “shark’s fin” appearance (Figure 5a). The steeper phase III also causes the end-tidal value to depend more strongly on expiratory time. This feature, combined with the anatomic dead space inherent in emphysema, causes marked discrepancies between end-tidal and arterial PCO2 values. These qualitative features of obstructive diseases on time-capnography correspond to quantitative aspects on volume-capnography. An increased phase III slope has been demonstrated in most diseases with airflow obstruction (COPD, asthma and bronchiectasis) 36, 37. The degree of slope correlates with the severity of airflow obstruction measured by spirometry36,37,38,39,40,41 and the degree of emphysema seen on chest CT 42. Automated analysis of multiple capnographic indices may increase the sensitivity and specificity for diagnosing obstructive lung disease. Combining exhalation duration, ET-CO2, endexhalation slope, and time spent at ET-CO2, succeeds in distinguishing patients with COPD from those
ACCEPTED MANUSCRIPT
with congestive heart failure.43 Capnography may be even more sensitive than spirometry for detecting early, small airway changes, as has also been demonstrated for the similar technique of single-breath nitrogen washout.44,45
SC
RI PT
ET-CO2 may differ markedly from the simultaneous PaCO2, especially in obstructed patients, and this is true when patients are stable or having an exacerbation.46,47 A slow, forced, maximal expiration can produce a much better correlation between ET-CO2 and PaCO246, but this is not practical in the critically ill. In some ventilated patients with COPD the difference between ET-CO2 and PaCO2 can exceed 40mmHg, emphasizing the serious errors that can be made in drawing inferences about the adequacy of ventilation based on ET-CO2 values.
Capnography and cardiac output:
It is commonly recognized that ET-CO2 values change with the return of spontaneous circulation (ROSC) ; when passive leg raising (PLR) augments cardiac output51, 52, 53; and in many shock states54, 55. These findings give rise to the incorrect idea that cardiac output directly influences ET-CO2. In steady state, however, the volume of CO2 exhaled through the lungs must equal that produced in the tissues metabolically, regardless of cardiac output56. Cardiac output largely affects ET-CO2 only in dynamic situations, not when the circulation is stable.
M AN U
48,49, 50
EP
TE D
The appearance of a direct relationship between cardiac output and ET-CO2 relates to non-steady state conditions combined with the fact that the body can store large amounts of CO2.55 57 58For example, if cardiac output is reduced through hemorrhagic shock, ET-CO2 falls transiently as less CO2 returns to the lungs. At the lower cardiac output, tissue and venous PCO2 rise, ultimately restoring the total volume of CO2 reaching the lungs, balancing metabolic production with excretion. Because so much CO2 can be stored in the body as tissue CO2 rises, equilibrium will not be reached quickly. In addition, many interventions to change cardiac output also change dead space (for example, by raising or lowering left atrial pressure).59 60 Thus short-term experiments may give the appearance of a direct relationship of cardiac output and time-capnography.
AC C
There is a more subtle effect of cardiac output on alveolar and ET-CO2 related to the Fick Principle. Recalling that alveolar CO2 tends to approximate the venous value at end-expiration, it should be clear that low cardiac output states will produce larger (but still modest) swings in alveolar and ET-CO 2. The tendency, then, is for low cardiac output to produce higher (not lower) ET-CO2 in steady-state and when all other variables are kept constant. Although ET-CO2 does not directly reflect cardiac output, by signaling changes, it has great value during resuscitation and when judging fluid-responsiveness. Cardiopulmonary resuscitation (CPR): During CPR of cardiac arrest, capnography signifies increasing pulmonary blood flow, thus the adequacy of chest compressions.61 62 An ET-CO2 value of
