Carbon Dioxide Production, Transport, and Elimination

An arterial carbon dioxide value is the result of a moving system. Tissues generate carbon dioxide, blood transports it in several forms, and the lungs eliminate it through effective ventilation. Changes in metabolism, circulation, storage, or gas exchange can therefore produce similar blood-gas findings through different mechanisms.

35.1 Carbon Dioxide Production and Elimination Are Different Rates

Cells generate carbon dioxide during oxidation of metabolic fuels. Carbon released from these reactions reaches blood and is ultimately exhaled. The amount generated per minute depends on the total metabolic activity and on the substrates being used.

Production is not identical to the carbon dioxide measured at the airway each minute. Blood, extracellular fluid, and tissues contain substantial exchangeable carbon dioxide, mainly in chemically converted forms. When these stores change, pulmonary elimination can temporarily differ from tissue production.

A simplified whole-body balance makes the distinction explicit:

For ordinary steady-state interpretation, metabolic generation is the main input and pulmonary excretion is the main output. Exogenous carbon dioxide, extracorporeal removal, and other important pathways must be added when present. Conversions between bicarbonate and dissolved carbon dioxide redistribute the stored forms and can change what is delivered to the lungs.

After an abrupt ventilation increase, the lungs may initially excrete carbon dioxide accumulated before the change. Measured excretion can exceed current metabolic production while stores are being depleted. After ventilation decreases, elimination may initially fall below production and stores rise until a new balance develops.

The compartments do not equilibrate at identical speeds. Alveolar gas and circulating blood respond relatively quickly, while tissue reservoirs exchange over longer intervals. A changing arterial PCO₂ therefore does not imply that every tissue has reached the same new carbon dioxide state.

Acid–base changes add another influence. Buffering a nonvolatile acid with bicarbonate can increase carbon dioxide presented for elimination without representing a matching increase in oxidative metabolism. Similarly, observing more expired carbon dioxide after circulation improves may reflect washout of previously retained carbon dioxide as well as new production.

This is why one expired measurement should not be treated as a direct metabolic rate during an unstable transition. The relationship becomes more useful when ventilation, circulation, and storage are sufficiently stable. State whether a value describes tissue production, delivery to the lung, or measured pulmonary elimination.

35.2 Metabolic Demand and Nutritional Supply Influence the Load

The respiratory quotient is the ratio of metabolic carbon dioxide production to oxygen consumption. Different fuels produce different ratios during oxidation. Carbohydrate oxidation generally produces more carbon dioxide per unit oxygen consumed than fat oxidation, but the clinical load also depends on how much total energy is being metabolized.

The respiratory exchange ratio measured at the airway is the ratio of carbon dioxide elimination to oxygen uptake. It approximates the metabolic quotient under suitable steady-state conditions. During store changes, buffering, or unstable gas exchange, the measured ratio can depart from the underlying substrate-oxidation ratio.

Excess energy delivery can increase carbon dioxide generation and the ventilation needed to excrete it. Changing the carbohydrate fraction is therefore not the only relevant nutritional variable. Energy intake, endogenous metabolism, substrate handling, and measurement conditions all matter.

Talpers and colleagues compared different parenteral nutritional regimens in stable mechanically ventilated patients. Within their studied conditions, increasing total caloric delivery increased carbon dioxide production more clearly than changing the carbohydrate proportion of isocaloric regimens. This physiological study supports examining total energy supply; it does not establish one nutritional prescription for every critically ill patient. 1

Fever, agitation, shivering, physical activity, and increased respiratory-muscle work can also raise metabolic demand. The resulting carbon dioxide load can increase the ventilation requirement even if lung mechanics and dead-space fraction are unchanged. Conversely, reduced activity or temperature can change the load independently of lung recovery.

Manthous and colleagues measured metabolic responses during cooling of febrile critically ill patients. Their observations demonstrate that temperature-related changes can affect oxygen consumption and carbon dioxide handling. They do not establish that routine temperature reduction improves clinical outcomes, and the metabolic effect of cooling can differ if it provokes shivering. 2

A rising PaCO₂ should therefore prompt assessment of the load as well as the ventilator. Increasing ventilation may compensate for greater production, but the additional mechanical work or ventilator exposure has consequences. Identifying a reversible increase in demand can explain why the prior support no longer maintains the same carbon dioxide level.

35.3 Dissolved Gas, Bicarbonate, and Carbamino Compounds Carry Carbon Dioxide

Carbon dioxide leaving tissues dissolves in blood and diffuses into red cells. Some remains physically dissolved. Much is converted to bicarbonate, and some binds to amino groups on proteins, especially hemoglobin, forming carbamino compounds.

These forms are linked by reversible reactions rather than separate transport channels that never interact. Removing dissolved carbon dioxide in the lung draws carbon dioxide out of the chemically stored forms. Adding carbon dioxide in tissues shifts the balance in the opposite direction.

Carbonic anhydrase greatly accelerates the hydration and dehydration steps. Without rapid catalysis, the uncatalyzed reaction would be too slow to support normal exchange during brief capillary transit. Meldrum and Roughton's experimental work characterized this enzyme and its properties, establishing a key mechanism of rapid blood carbon dioxide handling. 3

Hydrogen ions generated during bicarbonate formation are buffered by hemoglobin and other buffers. Buffering limits the immediate rise in free hydrogen-ion concentration and allows more carbon dioxide to be carried in converted form. It does not mean that carbon dioxide loading has no acid–base consequence.

Carbamino binding is chemically distinct from oxygen binding to heme. Carbon dioxide reacts with suitable amino groups, while oxygen binds reversibly at the heme iron. Changes in hemoglobin's oxygenation state influence its carbon dioxide and hydrogen-ion binding, linking the two transport systems without making their binding sites identical.

The fractions carried in each form vary with oxygenation, pH, hemoglobin, and the comparison being made. The fraction of total blood carbon dioxide present as bicarbonate is not necessarily the same as bicarbonate's contribution to the small arteriovenous increment carried from tissues during one passage. Fixed percentages can obscure that distinction.

Blood PCO₂ reflects dissolved-gas tension, not the sum of all transported forms. A routine chemistry measurement labeled total CO₂ is also different from PCO₂ and consists mainly of bicarbonate in the sampled serum or plasma. Neither label should be used interchangeably with whole-blood carbon dioxide content without defining the sample and units.

35.4 Red-Cell Bicarbonate Exchange Extends Transport into Plasma

Much bicarbonate is generated inside red cells because they contain abundant carbonic anhydrase and hemoglobin buffer. The bicarbonate does not all remain inside the cell. It can move into plasma while chloride moves in the opposite direction through the erythrocyte anion exchanger, commonly called band 3 or AE1.

This chloride shift helps maintain electrical balance during exchange. As tissue carbon dioxide enters the red cell, hydration produces bicarbonate and hydrogen ions. Hemoglobin buffers much of the hydrogen load. Bicarbonate can then leave in exchange for chloride, allowing the plasma compartment to carry a substantial part of the transported carbon dioxide in bicarbonate form.

At the lung, the sequence reverses. Bicarbonate enters the red cell as chloride leaves. Bicarbonate combines with hydrogen ions, and carbonic anhydrase accelerates conversion back to carbon dioxide and water. Carbon dioxide then diffuses toward alveolar gas and is exhaled.

The reactions and membrane exchange operate together. Their direction reflects local gradients and the changing chemical state of hemoglobin. The chloride movement is not the source of metabolic carbon dioxide and does not require a net chloride gain by the body during each circuit.

Swietach and colleagues directly measured intracellular pH in individual human red cells. Recovery from acid and base loads depended on chloride/bicarbonate exchange through AE1. Their cellular experiments support this exchange mechanism; they do not make a plasma chloride concentration a direct measurement of carbon dioxide transport speed. 4

The red cell also carries oxygen in the opposite overall direction. Oxygen loading in pulmonary capillaries changes hemoglobin's buffering and carbon dioxide-binding behavior, facilitating unloading. Deoxygenation during tissue transit supports the reverse process. This coupling helps blood exchange both gases efficiently within the limited time available.

The immediate transport system should be distinguished from renal adaptation. Red-cell bicarbonate–chloride exchange redistributes ions between cell and plasma on the circulatory time scale. Renal bicarbonate conservation and net acid excretion alter whole-body acid–base balance over a longer period.

35.5 Oxygenation Changes the Carbon Dioxide Dissociation Curve

The carbon dioxide dissociation curve relates whole-blood carbon dioxide content to PCO₂ under specified conditions. It is not one immutable curve. Hemoglobin concentration, oxygenation, pH, and temperature influence how much total carbon dioxide blood carries at a given tension.

Deoxygenated blood can carry more carbon dioxide at the same PCO₂ than oxygenated blood. Deoxygenated hemoglobin binds hydrogen ions more effectively and supports additional bicarbonate formation. Its carbamino-binding properties also contribute. This oxygen-dependent change in carbon dioxide carriage is the Haldane effect.

Christiansen, Douglas, and Haldane studied the absorption and dissociation of carbon dioxide by human blood and demonstrated the influence of oxygenation. Their work established the central physiological relationship. The later mechanistic account includes several linked chemical processes rather than treating the observation as only one binding reaction. 5

The Haldane effect differs from the Bohr effect. The Haldane effect describes how oxygenation changes carbon dioxide and hydrogen-ion carriage. The Bohr effect describes how carbon dioxide and hydrogen-ion conditions change hemoglobin's oxygen affinity. The processes interact, but naming one does not identify the other.

Oxygenating venous blood therefore promotes carbon dioxide unloading in the lung. If the released carbon dioxide cannot be eliminated adequately, PCO₂ can rise. This is one contributor to oxygen-related carbon dioxide changes in susceptible patients; it does not account for every effect of oxygen, which can also alter V/Q matching and ventilation.

At a fixed PCO₂, a lower pH reduces the bicarbonate component relative to conditions with a higher pH. Temperature changes alter solubility and the chemical equilibria. Warming a fixed blood sample can raise its gas tension without adding an equivalent amount of new carbon dioxide. Interpret values with the temperature convention used by the analyzer.

Douglas and colleagues evaluated equations for calculating whole-blood carbon dioxide content using measured blood variables. Their findings support including oxygenation and other conditions rather than converting pressure to content with one fixed factor. A formula derived under defined conditions should not be treated as an error-free content assay in every ICU state. 6

35.6 A Venous–Arterial Carbon Dioxide Gap Is Not a Direct Flow Measurement

Blood flowing through tissues acquires carbon dioxide. Under a steady whole-body balance, the amount transported per minute equals blood flow multiplied by the mixed venous-to-arterial carbon dioxide content difference. This is the carbon dioxide form of the Fick balance.

Let V̇CO₂ be net carbon dioxide transfer in mL/min, Q cardiac output in L/min, CvCO₂ mixed venous content, and CaCO₂ arterial content, both in mL CO₂/dL. The factor 10 converts liters to deciliters.

For an illustrative transfer of 200 mL/min at an output of 5 L/min, the required content difference is 200 / (5 × 10) = 4 mL CO₂/dL. At 2.5 L/min, unchanged transfer would require 200 / 25 = 8 mL CO₂/dL. Lower flow requires a greater content increment per unit blood when the other assumptions hold.

The commonly measured clinical gap uses PCO₂ rather than content. Because the dissociation relationship changes with oxygenation, pH, temperature, and hemoglobin, that pressure gap cannot be substituted directly into the content equation. A difference of 8 mm Hg is not the same quantity as 8 mL CO₂/dL.

Vallet and colleagues compared regional low-flow ischemic hypoxia with hypoxia caused by reduced arterial oxygen in an experimental preparation. The carbon dioxide pressure difference behaved differently between these mechanisms. Their findings show why a gap is not a universal detector of inadequate tissue oxygenation independent of flow. 7

Cuschieri and colleagues examined central and mixed venous gaps alongside cardiac-index measurements in ICU patients. This prospective cohort supports studying the gap as contextual flow-related information, while central venous and mixed venous samples remain different sampling compartments. 8

In septic shock, Ospina-Tascón and colleagues associated persistently high gaps with organ dysfunction and worse outcomes, but agreement with cardiac output was poor across observations. The prognostic finding does not establish that normalizing the gap by fluids or another intervention improves survival. Interpret production, transport chemistry, sampling, and the rest of perfusion together. 9

35.7 Arterial PCO₂ Depends on Effective Alveolar Ventilation

At steady state, alveolar ventilation must remove the carbon dioxide delivered for pulmonary elimination. A greater elimination rate requires more ventilation to maintain the same alveolar carbon dioxide tension. For a given elimination rate, less effective alveolar ventilation produces a higher tension.

Let V̇CO₂ be pulmonary carbon dioxide elimination expressed in mL/min at standard temperature and pressure, dry, or STPD. Let V̇A be effective alveolar ventilation in L/min at body temperature and ambient pressure, saturated, or BTPS. With PCO₂ expressed in mm Hg, a commonly used relationship is:

The arterial-to-alveolar approximation is most appropriate when the model's ideal gas-exchanging compartment represents the blood–gas relationship adequately. Marked V/Q inequality and shunt complicate that interpretation. Inspired carbon dioxide is assumed negligible, and a sufficiently steady state is required.

For an illustrative elimination rate of 200 mL/min and effective alveolar ventilation of 4 L/min, PCO₂ is approximately 0.863 × 200 / 4 = 43.2 mm Hg. If elimination remains 200 but effective ventilation falls to 3 L/min, the value becomes approximately 57.5 mm Hg. If elimination instead rises to 300 while ventilation remains 4, the value becomes approximately 64.7 mm Hg.

The examples show two routes to hypercapnia: less effective ventilation or a greater carbon dioxide load relative to that ventilation. They are separate mechanisms even when the resulting arterial value is similar.

If V̇CO₂ is entered in L/min rather than mL/min, the coefficient is 863 with V̇A still in L/min. Thus 863 × 0.200 / 4 gives the same approximately 43.2 mm Hg. Mixing a coefficient with the wrong volume units creates a thousand-fold error.

In true steady state, pulmonary elimination approximates production when other gains, losses, and store changes are negligible. During an abrupt circulatory or ventilatory change, that equivalence may fail. The equation should not turn a transient airway elimination value into a direct estimate of ongoing metabolic production.

35.8 Total Minute Ventilation Can Remain Stable While Elimination Worsens

Minute ventilation is respiratory rate multiplied by tidal volume. It includes gas that does not contribute effectively to carbon dioxide clearance. A stable displayed minute ventilation therefore does not establish stable effective alveolar ventilation.

In a simple compartment model, effective alveolar ventilation is rate multiplied by tidal volume minus dead-space volume. Let rate be 20 breaths/min, tidal volume 500 mL, and effective dead space 200 mL per breath. Each breath contributes 300 mL to the modeled effective ventilation, giving 20 × 300 = 6,000 mL/min, or 6 L/min.

If effective dead space rises to 300 mL while rate and tidal volume remain unchanged, effective ventilation falls to 20 × 200 = 4 L/min. Total minute ventilation remains 10 L/min in both cases. The resulting carbon dioxide change can therefore occur without a decrease in the volume displayed by the ventilator.

The example uses an equivalent compartment model rather than a direct anatomical measurement. Different dead-space methods, especially Bohr and Enghoff approaches, describe different combinations of ventilatory dilution and gas-exchange inequality. An Enghoff-derived fraction should not automatically be treated as a literal volume of unperfused alveoli.

Apparatus can also add rebreathed volume. The same added common-path volume consumes a larger fraction of a small tidal breath. Circuit configuration, heat-and-moisture exchangers, connectors, leaks, and inspired carbon dioxide should therefore be considered when ventilation becomes unexpectedly ineffective.

Pulmonary vascular or cardiac changes can alter clearance independently of delivered volume. Reduced flow to ventilated lung can lower expired carbon dioxide initially while tissues retain more. Recruitment, overdistension, and redistribution can change the effectiveness of the same ventilatory pattern.

A compensatory increase in rate or volume may improve elimination but can also increase mechanical exposure, trapping, or cardiovascular loading. Evaluate the cause of inefficiency and the full response rather than assuming that restoring one arterial PCO₂ justifies any increase in support.

35.9 Hypercapnia Changes Acid–Base Balance and Organ Conditions

Rising carbon dioxide shifts the hydration reaction toward hydrogen ions and bicarbonate. The immediate result tends toward respiratory acidemia, moderated by nonbicarbonate buffers. The same PaCO₂ can accompany different pH values depending on baseline bicarbonate, prior adaptation, and concurrent metabolic disturbances.

Over a longer period, functioning kidneys increase bicarbonate conservation and net acid excretion, helping restore the bicarbonate-to-dissolved-carbon-dioxide relationship. Polak and colleagues studied adaptation to sustained hypercapnia in dogs using balance measurements. Their experiments established important features of the renal and electrolyte response, but do not provide an exact compensation rate for every human patient with kidney injury or mixed acid–base disease. 10

Adaptation does not remove the underlying carbon dioxide load or make the original ventilatory limitation disappear. Kidney dysfunction, chloride and potassium conditions, medications, and simultaneous metabolic acid production can change the response. A relatively preserved pH should therefore be interpreted with the trajectory and accompanying chemistry.

Carbon dioxide readily crosses the blood–brain barrier. A rise can acidify the perivascular environment, dilate cerebral arterioles, and increase cerebral blood flow and blood volume. If intracranial compliance is limited, the additional volume can raise intracranial pressure. Kety and Schmidt directly studied altered gas tensions in healthy young men and demonstrated substantial cerebral circulatory responses to carbon dioxide. The findings establish physiological sensitivity; they do not supply a safe PaCO₂ target for an injured brain with impaired compliance or altered vascular responsiveness. 11

Cardiovascular effects depend on both carbon dioxide and the associated acid–base state. Changes in vascular tone, sympathetic activity, myocardial function, and pulmonary vascular load can interact. A patient with limited right ventricular reserve may respond differently from one with preserved function.

Mekontso Dessap and colleagues studied ventilatory strategies that produced acute hypercapnia and acidemia while changing PEEP and tidal volume in severe ARDS. Right ventricular function and cardiac index deteriorated despite improved oxygenation and recruitment. Because several physiological conditions changed together, the study cannot assign the entire effect to PaCO₂ alone. 12

These distinctions prevent one universal safe hypercapnia threshold. Rate of change, pH, cerebral conditions, pulmonary vascular load, and the reason for limiting ventilation influence tolerance. Accepting a higher PaCO₂ to avoid excessive ventilation is also different from evidence that deliberately producing hypercapnia improves clinical outcomes.

35.10 Expired Carbon Dioxide Is a Continuous Signal With Several Determinants

Capnography shows the carbon dioxide reaching the measurement site during expiration. End-tidal PCO₂ samples the end of the expired waveform. Volumetric capnography adds expired-volume information and can estimate carbon dioxide elimination from the integral of concentration over volume.

Elimination is a volume-per-time quantity. End-tidal PCO₂ is a pressure. A changed end-tidal value does not by itself quantify the change in total carbon dioxide excretion, metabolic production, or arterial PCO₂. The waveform and breathing pattern influence its meaning.

An abrupt fall can reflect increased effective ventilation, reduced delivery of carbon dioxide by pulmonary blood flow, a leak, disconnection, or sampling failure. Complete loss of a previously clear signal requires immediate assessment of the patient and circuit. A displayed value without an adequate waveform is less secure evidence than a consistent measured breath.

Shibutani and colleagues studied acute hemodynamic changes during surgery under constant ventilation. Decreases in end-tidal carbon dioxide and measured elimination tracked decreases in cardiac output in that selected setting. This supports considering perfusion when the signal falls unexpectedly, but it does not establish a universal conversion from an end-tidal change to an output change in all ICU states. 13

A rise can reflect increased production or washout, reduced ventilation, increased rebreathing, or a changed emptying pattern. Restored circulation may increase delivery of stored tissue carbon dioxide before a stable new balance develops. The same numerical rise can therefore accompany improvement or deterioration through different mechanisms.

Compare the signal with actual volume delivery, respiratory rate, effort, hemodynamics, temperature, and recent interventions. Pair an arterial gas with the expired measurement when the arterial relationship matters. A previous arterial-to-end-tidal gap should not be assumed unchanged after a major change in perfusion or lung conditions.

Serial interpretation is strongest when acquisition and conditions are comparable. Use the continuous signal to detect a change early, then determine whether the cause lies in production, transport, ventilation, or measurement. That preserves its value without asking it to answer every physiological question alone.

35.11 Integrating a Rising Arterial Carbon Dioxide

Consider an illustrative ventilated patient whose PaCO₂ rises while total minute ventilation remains similar. Temperature has increased, an additional airway connector was added, and circulation has become less stable. The unchanged ventilator display does not isolate the mechanism.

Begin with the measurements. Confirm the arterial sample, waveform, delivered volumes, leak, circuit configuration, and inspired carbon dioxide when relevant. Determine whether the new connector added rebreathed volume and whether the patient is triggering extra or ineffective efforts.

Next assess load and effective ventilation separately. Increased metabolic activity can raise production. Additional apparatus dead space can reduce the useful portion of each breath. Pulmonary vascular changes or reduced output can alter the distribution and time course of elimination. More than one process may be present.

The expired signal helps interpret the timing. A fall in end-tidal carbon dioxide during circulatory deterioration may indicate reduced delivery to the lung despite rising tissue and arterial carbon dioxide. A later rise after circulation improves may partly reflect washout. Neither change should be attributed automatically to a new metabolic rate.

Review pH and the prior course as well. A sudden PaCO₂ rise without established adaptation differs from a similar value during chronic hypercapnia. Concurrent metabolic acidosis or kidney dysfunction changes the acid–base consequence. Cerebral vulnerability and right ventricular function influence how the new state is tolerated.

If venous-to-arterial PCO₂ data are available, interpret the gap as additional contextual information. Sampling site, oxygenation, hemoglobin, pH, and temperature affect the relation between pressure and transported content. A large gap can strengthen concern about inadequate transport without independently defining a fluid requirement or proving anaerobic metabolism.

The useful explanation connects production, storage, blood transport, pulmonary perfusion, and effective ventilation over time. An intervention can then be judged by the circulation, gas exchange, effort, and organ effects it produces. The objective is an adequate integrated state, not merely a familiar PaCO₂ obtained at any mechanical or systemic consequence.

References

1. Talpers SS, Romberger DJ, Bunce SB, Pingleton SK. Nutritionally associated increased carbon dioxide production. Excess total calories vs high proportion of carbohydrate calories. Chest. 1992;102(2):551–555. doi:10.1378/chest.102.2.551

2. Manthous CA, Hall JB, Olson D, et al. Effect of cooling on oxygen consumption in febrile critically ill patients. Am J Respir Crit Care Med. 1995;151(1):10–14. doi:10.1164/ajrccm.151.1.7812538

3. Meldrum NU, Roughton FJW. Carbonic anhydrase. Its preparation and properties. J Physiol. 1933;80(2):113–142. doi:10.1113/jphysiol.1933.sp003077

4. Swietach P, Tiffert T, Mauritz JMA, et al. Hydrogen ion dynamics in human red blood cells. J Physiol. 2010;588(Pt 24):4995–5014. doi:10.1113/jphysiol.2010.197392

5. Christiansen J, Douglas CG, Haldane JS. The absorption and dissociation of carbon dioxide by human blood. J Physiol. 1914;48(4):244–271. doi:10.1113/jphysiol.1914.sp001659

6. Douglas AR, Jones NL, Reed JW. Calculation of whole blood CO₂ content. J Appl Physiol. 1988;65(1):473–477. doi:10.1152/jappl.1988.65.1.473

7. Vallet B, Teboul JL, Cain S, Curtis S. Venoarterial CO₂ difference during regional ischemic or hypoxic hypoxia. J Appl Physiol. 2000;89(4):1317–1321. doi:10.1152/jappl.2000.89.4.1317

8. Cuschieri J, Rivers EP, Donnino MW, et al. Central venous-arterial carbon dioxide difference as an indicator of cardiac index. Intensive Care Med. 2005;31(6):818–822. doi:10.1007/s00134-005-2602-8

9. Ospina-Tascón GA, Bautista-Rincón DF, Umaña M, et al. Persistently high venous-to-arterial carbon dioxide differences during early resuscitation are associated with poor outcomes in septic shock. Crit Care. 2013;17:R294. doi:10.1186/cc13160

10. Polak A, Haynie GD, Hays RM, Schwartz WB. Effects of chronic hypercapnia on electrolyte and acid-base equilibrium. I. Adaptation. J Clin Invest. 1961;40(7):1223–1237. doi:10.1172/JCI104353

11. Kety SS, Schmidt CF. The effects of altered arterial tensions of carbon dioxide and oxygen on cerebral blood flow and cerebral oxygen consumption of normal young men. J Clin Invest. 1948;27(4):484–492. doi:10.1172/JCI101995

12. Mekontso Dessap A, Charron C, Devaquet J, et al. Impact of acute hypercapnia and augmented positive end-expiratory pressure on right ventricle function in severe acute respiratory distress syndrome. Intensive Care Med. 2009;35(11):1850–1858. doi:10.1007/s00134-009-1569-2

13. Shibutani K, Muraoka M, Shirasaki S, Kubal K, Sanchala VT, Gupte P. Do changes in end-tidal PCO₂ quantitatively reflect changes in cardiac output? Anesth Analg. 1994;79(5):829–833. doi:10.1213/00000539-199411000-00002