Show Summary Details
Page of

Gas exchange principles in the critically ill 

Gas exchange principles in the critically ill
Chapter:
Gas exchange principles in the critically ill
Author(s):

Peter D. Wagner

DOI:
10.1093/med/9780199600830.003.0075
Page of

PRINTED FROM OXFORD MEDICINE ONLINE (www.oxfordmedicine.com). © Oxford University Press, 2020. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

date: 01 December 2020

Key points

  • Gas exchange is a process of molecular movement governed by the unifying principle of mass conservation applied equally to both gas uptake and elimination.

  • Pulmonary gas exchange in the critically-ill patient is almost always impaired with multiple causes both internal and external to the lung.

  • In a normal subject breathing sea level air at rest, the V ˙ A / Q ˙ ratio of the lung as a whole is around 1 because total alveolar ventilation and pulmonary blood flow are about the same. In the critically ill, inequality can be very severe.

  • The V ˙ A / Q ˙ ratio uniquely determines PaO2 and PaCO2 in a given lung region, since all lung regions are subject to the same boundary conditions and O2-Hb and CO2 dissociation curves.

  • Hypoxaemia and hypercapnia are caused by V ˙ A / Q ˙ inequality, inspiratory hypoxia, hypoventilation, diffusion limitation, shunt, or extrapulmonary factors.

Introduction

This chapter discusses the several reasons for, and mechanisms of, abnormal gas exchange in critically-ill patients in intensive care units (ICUs). The principles underlying gas exchange have been well known since the 1950s when Rahn and Fenn [1]‌, and Riley [2,3,4] and colleagues developed quantitative approaches to its understanding. While these principles apply equally to normal subjects and patients with lung disease, nowhere is their application and understanding both more difficult and also more important than in the ICU. It is well known that patients in the ICU usually endure severe lung disease, no matter what the cause of their illness, and that supportive interventions to improve gas exchange carry their own, sometimes severe, risks of further lung damage.

What follows applies only under steady-state conditions—that is, conditions in which concentrations of gases (such as O2 and CO2) are constant in time for at least several minutes. Rapid changes in cardiopulmonary structure or function frequently occur in the ICU, and then the steady state is disturbed, with O2 and CO2 concentrations rapidly changing, and the use of steady-state concepts will be in error.

Basic principle: mass conservation

Gas exchange is a process of molecular movement governed by the unifying principle of mass conservation, both when movement is by convection (ventilation and blood flow) and by diffusion (gas transfer across the alveolar-capillary blood gas barrier). It applies equally to gas uptake (e.g. O2) and elimination (e.g. CO2).

To best understand this, one applies conservation of mass equations, similar for O2 and CO2, so that illustrating the case using O2 is here sufficient. Conservation of mass decrees that all the inhaled O2 not exhaled on the next breath (eqn 1) is transferred into the pulmonary capillary blood (eqn 2). This transfer rate is the O2 uptake, V ˙ O 2 :

V˙ O 2 = V˙ A [ F I O 2     F A O 2 ]
[eqn 1]

V˙ O 2   =   Q˙   [ CaO 2     CvO 2 ]
[eqn 2]

Taking eqns 1 and 2, and rearranging gives:

V ˙ A / Q ˙ = [ C a O 2 C v O 2 ] / [ F I O 2 F A O 2 ]
[eqn 3]

V ˙ A is alveolar ventilation and Q ˙ total pulmonary blood flow; CaO2 and CvO2 are systemic and pulmonary arterial O2 concentrations, respectively; FiO2 and FaO2 are inspired and alveolar O2 concentrations respectively. A minor approximation appears above—inspired and expired gas volumes are taken to be the same, which is true to within about 1%.

This equation is written for the whole lung, so that CaO2 is mixed arterial blood [O2], while FaO2 is mixed alveolar gas O2 fractional concentration. It applies equally to small homogeneous regions in the lung (i.e. approximately, the acinus [5]‌, but the terminology differs:

V˙ A / Q˙   =   [ Cc O 2 CvO 2 ] / [ F I O 2 F A O 2 ]
[eqn 4]

where ‘c'’, or ‘end-capillary’, represents [O2] in the pulmonary venule draining the homogeneous region, and FaO2 is now the alveolar [O2] in the same region.

Eqn (4) shows that three factors determine gas exchange:

  • The region’s V ˙ A / Q ˙ ratio.

  • The ‘boundary conditions’, i.e. O2 composition of inspired gas (FiO2) and pulmonary arterial blood (CvO2).

  • The characteristics of the O2-Hb dissociation curve.

The latter is involved because Cc’O2 and FaO2 are directly linked through that curve. Since by Dalton’s law of partial pressures PO2 = FO2 * [Pb–PH2O], the alveolar PO2 (PaO2) is also defined. Thus, there is only a single value of PaO2 that satisfies eqn 4 for any set of the three groups of variables listed immediately previously. This is shown in Fig. 75.1, where the boundary conditions and O2Hb dissociation curve are held constant (using normal values), while the V ˙ A / Q ˙ ratio is systematically varied from zero to infinity. Fig. 75.1 also shows the results for CO2, based on identical considerations and normal values. Eqn 4 is called the ventilation/perfusion ( V ˙ A / Q ˙ ) eqn for O2. For CO2, it is in essence identical, but because CO2 is eliminated while O2 is taken up, the quantities appear in reverse order to avoid negative signs:

Fig. 75.1 Semi-logarithmic plot of alveolar gas PO2 and PCO2 as a function of the ventilation/perfusion ratio.

Fig. 75.1 Semi-logarithmic plot of alveolar gas PO2 and PCO2 as a function of the ventilation/perfusion ratio.

V ˙ A   /   Q ˙   =   [ C v C O 2     C c C O 2 ] / [ F A C O 2     F I C O 2 ]
[eqn 5]

It is also useful to write the analog of eqn 1 for CO2, giving:

V˙ CO 2 = V˙ A * F A CO 2   V˙ I * F I CO 2
[eqn 6]

This is useful, because, disregarding inspired [CO2] as negligible, the ratio of eqn 6 to eqn 1, gives:

V˙ CO 2 / V˙ O 2 =   ( V˙ A   *   [ F A CO 2 ] )   / ( V˙ A   *   [ F I O 2     F A O 2 ] )
[eqn 7]

or,

V˙ CO 2   /   V˙ O 2   =   R   =   ( P A CO 2 ) / [ P I O 2     P A O 2 ]
[eqn 8]

because F (fractional alveolar gas concentration) and P (alveolar partial pressure) are proportional. V ˙ A cancels out of the equation; R is the respiratory exchange ratio (the ratio of CO2 produced to O2 consumed). Re-arranging gives:

P A O 2   =   P I O 2     P A CO 2 / R
[eqn 9]

This is the simple form of the alveolar gas equation. It suffices for clinical use, but if one wishes to account for V ˙ I and V ˙ A being slightly different, it becomes:

P A O 2   =   P I O 2     ( P A CO 2 / R) +   ( P A CO 2   * F I O 2   *   ( 1     R )   /   R)
[eqn 10]

The alveolar gas equation defines the PaO2 if we know PaCO2, PiO2, and R. It applies to the lung as a whole or to a small homogenous region, just like the ventilation/perfusion equation.

The most important outcome is understanding that the V ˙ A / Q ˙ ratio uniquely determines PaO2 and PaCO2 in a given lung region (since all lung regions are subject to the same boundary conditions and O2-Hb and CO2 dissociation curves); see Fig. 75.1.

Ventilation/perfusion matching and gas exchange

If all lung regions had the same V ˙ A / Q ˙ ratio, they would all have the same PaO2 and PaCO2. In a normal subject breathing sea level air at rest, the V ˙ A / Q ˙ ratio of the lung as a whole is around 1 because total alveolar ventilation and pulmonary blood flow are about the same at 5–6 L/min. PaO2 is ~100 mmHg, while PaCO2 is ~40 mmHg (Fig. 75.1). Because all regions reach diffusion equilibrium of O2 exchange across the blood–gas barrier at rest in health [6]‌, end capillary PO2 and PCO2 would equal their alveolar counterparts in each region, and thus the systemic arterial blood would also have a PO2 of 100 mmHg and PCO2 of 40 mmHg. However, if regional V ˙ A / Q ˙ ratio was lower than average. Fig. 75.1 shows PaO2 is reduced and PaCO2 increased in that region. Symmetrically, in a high V ˙ A / Q ˙ region, PaO2 will rise, and PACO2 fall.

Using Fig. 75.1, we may consider what we call V ˙ A / Q ˙ inequality, when not all regions have the same values of V ˙ A / Q ˙ . In the critically ill, inequality can be very severe. A simple ‘two-compartment’ model suffices to illustrate the principles, lung is usually far more complex. The top panel (Fig. 75.2A) shows the two-compartment model in the homogeneous state, with both ventilation ( V ˙ A ) and blood flow ( Q ˙ ) equally distributed. With total ventilation and perfusion each set to 6 L/min, the V ˙ A / Q ˙ ratio is 1.0 in each compartment. From Fig. 75.1, PaO2 is 104 mmHg and PaCO2 39 mmHg in each compartment. Mixed exhaled gas will also have these partial pressures, as will the mixed pulmonary venous blood returning to the left heart, and then systemic arterial PO2.

Fig. 75.2 (a) A two-compartment model of the lungs, with equally distributed ventilation and blood flow, resulting in a homogeneous lung. (b) A two-compartment model of the lungs with ventilation reduced by 90% on the left side (see text for description).
Fig. 75.2 (a) A two-compartment model of the lungs, with equally distributed ventilation and blood flow, resulting in a homogeneous lung. (b) A two-compartment model of the lungs with ventilation reduced by 90% on the left side (see text for description).

Fig. 75.2 (a) A two-compartment model of the lungs, with equally distributed ventilation and blood flow, resulting in a homogeneous lung. (b) A two-compartment model of the lungs with ventilation reduced by 90% on the left side (see text for description).

If the left airway is now largely partially obstructed (by mucus, bronchoconstriction, airway wall thickening, aspirated foreign object, tumour, etc.) as in the lower panel, Fig. 75.2B, V ˙ A of the obstructed compartment will be reduced (here by 90%) to 0.3 L/min. Because we wish to model only the effects of V ˙ A / Q ˙ inequality, total ventilation, and blood flow are unchanged. Thus, the right-hand compartment now receives more ventilation than normal, here 5.7 L/min. There is, thus, one compartment with low V ˙ A / Q ˙ ratio (0.1) and one with high V ˙ A / Q ˙ ratio (1.9), but the total ventilation and blood flow remain normal. From Fig. 75.1, the alveolar (and end capillary) PO2 are 45 (low V ˙ A / Q ˙ unit) and 122 mmHg (high V ˙ A / Q ˙ unit). PO2 in mixed exhaled gas (Pe) is the ventilation-weighted average of the two PaO2 values. Ventilation weighting is required to ensure conservation of mass.

P E   =   [ ( 0. 3 * 45 )   +   ( 5 . 7 * 122 ) ] / 6 .0   =   118   mmHg
[eqn 11]

This calculation should strictly use concentration, rather than partial pressure. However, partial pressure can be used because Dalton's Law states that partial pressure and concentration are proportional.

This is 14 mmHg higher than in the homogeneous lung (Fig. 75.2A). Inspired PO2 is about 150 mmHg and thus the inspired-exhaled PO2 difference in the homogeneous state is 150 – 104 = 46 mmHg and after airway obstruction is 150 – 118 = 32 mmHg. If mixed exhaled gas has a higher PO2, more O2 is exhaled and therefore less is transferred into the blood: V ˙ o 2 has fallen to 32/46 (i.e. ~70%) of that in the homogeneous lung.

PO2 in the mixed pulmonary venous blood is calculated from the perfusion-weighted average of the two O2 concentration values, using the O2-Hb dissociation curve to compute the corresponding PO2. This PO2 is only 58 mmHg.

For CO2, the numbers are quite different, but the outcome is qualitatively identical, but opposite: CO2 elimination is reduced, but only by ~13%, and arterial PCO2 rises, but by less than 1 mmHg. Indeed, the exchange of all gases is impaired by V ˙ A / Q ˙ inequality and is manifest by a fall in arterial blood values for all gases taken up (e.g. O2) and a rise in arterial blood values for all gases being eliminated (e.g. CO2).

In this example, the effects are greater for O2 than for CO2. The same analysis obstructing the pulmonary artery of the left compartment (and restoring equal ventilation distribution) is shown in Fig. 75.3. Here, everything else is identical to that in Fig. 75.2B. The resulting compartmental ratios of V ˙ A / Q ˙ are 10.0 and 0.53. As in the airway obstruction model, the two compartments have a below-average (right) and above-average (left) V ˙ A / Q ˙ ratio. Reading off Fig. 75.1, PaO2 are 142 and 77 mmHg, respectively. Mixed exhaled PO2 is 110 mmHg (weighted mean of 142 and 77), while arterial PO2 is 78 mmHg. V ˙ o 2 is reduced, to (150–110)/(150–104), or 87% of control, while V ˙ co 2 falls to 75% of control and arterial PCO2 increases by 2 mmHg.

Fig. 75.3 A two-compartment model of the lungs with blood flow reduced by 90% on the left side (see text for description).

Fig. 75.3 A two-compartment model of the lungs with blood flow reduced by 90% on the left side (see text for description).

Thus, the two models—airway and vascular obstruction—both show impaired overall gas transfer, hypoxaemia, and hypercapnia, but airway obstruction affects O2 transfer more than it affects CO2, while with vascular obstruction, CO2 transfer is more affected.

The causes of hypoxaemia and hypercapnia

There are six possible physiological causes of hypoxaemia and hypercapnia:

  • V ˙ A / Q ˙ inequality.

  • Inspiratory hypoxia.

  • Hypoventilation.

  • Diffusion limitation.

  • Shunt.

  • Extra-pulmonary factors related to metabolic rate and cardiac output.

V ˙ / Q ˙ inequality

Very important in critically-ill patients, inequality is defined as a range of V ˙ A / Q ˙ ratios throughout the lung. In some regions, the V ˙ A / Q ˙ ratio is low, in others normal, and in others high. This impairs overall O2 uptake and CO2 elimination (until compensated by increased O2 extraction, ventilation, or cardiac output), and causes hypoxaemia and hypercapnia.

There are many reasons for V ˙ A / Q ˙ inequality in ICU patients. They result from the pathological processes in the lung—infection, inflammation, fluid accumulation, vascular obstruction, tissue breakdown, and effects of ventilator management and other therapies. Importantly, these processes are never uniformly distributed throughout the lung. They are conveniently discussed by site of origin—airways, alveoli, blood vessels, and pleura. What follows is not exhaustive, but addresses the majority of mechanisms of inequality.

In the airways, bronchial wall thickening, mucus secretions, alveolar debris (cells, bacteria, inhaled particles, fluid) moved up the airways by ciliary function, bronchial smooth muscle contraction, and diminished parenchymal radial traction on the airway wall (the latter especially with pre-existing COPD) will each act to reduce airway lumen size, (partially) obstructing the airway, and increasing airflow resistance. Because some regions will be more obstructed than others, more affected regions become underventilated, and a low regional V ˙ A / Q ˙ ratio will usually develop. With overall lung ventilation maintained, less affected regions (taking the air originally designated for the obstructed regions) are over-ventilated, raising the V ˙ A / Q ˙ ratio above normal, and causing hyperinflation and risk of barotrauma.

In affected alveoli, cellular debris, oedema fluid, and bacterial fragments will (partly) displace the air, diminishing their ventilation (even if the airways are normal). Alveoli completely filled with such material will not be ventilated at all, resulting in shunt. Collapsed alveoli will not receive ventilation, also resulting in shunt. Impaired surfactant activity often occurs, reducing compliance and promoting atelectasis. Interstitial oedema will also reduce compliance, and, with airways factors mentioned in the preceding paragraph, reduce regional ventilation.

Thus, a large number of specific processes all acting to impair regional ventilation may be present in the lungs of critically-ill patients, creating regions of low V ˙ A / Q ˙ ratio, V ˙ A / Q ˙ inequality and hypoxaemia. At the same time, over-ventilation of less affected regions occurs. As pathological processes continue, partial airway obstruction may become complete, converting what was V ˙ A / Q ˙ inequality into shunt.

Pulmonary vascular pathology may also cause V ˙ A / Q ˙ inequality and hypoxaemia. In the ICU, one cause is pulmonary thromboembolism. It may be classical emboli from thrombosed veins in extra-pulmonary regions of the body or local thrombosis in damaged pulmonary vessels. The resulting (partial) vascular obstruction diverts blood flow to less affected lung regions. In this manner, the affected regions develop high V ˙ A / Q ˙ ratios due to loss of perfusion, while the unaffected regions become over-perfused, so their V ˙ A / Q ˙ ratio falls.

High V ˙ A / Q ˙ regions may develop from mechanical ventilation with high airway pressures. This may overventilate and hyperinflate the most normal, compliant alveoli, stretching alveolar walls, compressing alveolar capillaries [7]‌, increasing vascular resistance, reducing their perfusion and increasing the V ˙ A / Q ˙ ratio. At first sight, this may not seem problematic (a high V ˙ A / Q ˙ ratio corresponds to a high PO2 and low PCO2, Fig. 75.1). However, there are three major concerns:

  • Risk of barotrauma from the hyperinflation.

  • Poorly perfused alveoli are not effective for gas exchange, because little O2 can be added to the blood or CO2 removed.

  • The ventilation delivered to these alveoli causes the perfused alveoli that undertake most of the gas exchange to become under-ventilated and CO2 retention may develop.

Pleural processes that may affect gas exchange include pneumothorax and effusion, both of which promote atelectasis (and thus shunt), or pleural thickening that restricts lung inflation.

To summarize, many mechanisms may co-exist in ICU patients to produce V ˙ A / Q ˙ inequality. Those that primarily affect the airways, alveoli, and pleura lead mostly to regions of reduced V ˙ A / Q ˙ ratio, while those that primarily affect the vasculature result in regions of high V ˙ A / Q ˙ ratio. These different locations of effect may then result in different arterial PO2/PCO2 pictures, as expressed in Figs 75.2 and 75.3, gas exchange is always impaired.

Inspiratory hypoxia

Reduced inspired PO2 (e.g. ascent to altitude, plane flight) causes arterial hypoxaemia, but in the ICU, this should not occur and will not be discussed.

Hypoventilation

Insufficient ventilation to the lungs as a whole will cause arterial PO2 to fall and PCO2 to rise. Fig. 75.4 (representing eqns 1 and 6) shows how sensitive PO2 and PCO2 are to changes in ventilation, especially when reduced. Unless by therapeutic design, overall hypoventilation is not commonly seen in ICU patients.

Fig. 75.4 Alveolar PO2
 and PCO2 as a function of alveolar ventilation in a homogeneous lung. Vertical dashed line represent a normal ventilation, black dots represent the interception with PO2
 and PCO2 curves, and vertical dashed lines indicates the corresponding pressure values on the ordinate axis.

Fig. 75.4 Alveolar PO2 and PCO2 as a function of alveolar ventilation in a homogeneous lung. Vertical dashed line represent a normal ventilation, black dots represent the interception with PO2 and PCO2 curves, and vertical dashed lines indicates the corresponding pressure values on the ordinate axis.

Diffusion limitation

An assumption of the preceding analysis has been that diffusion of O2 and CO2 between alveolar gas and capillary blood is complete within the transit time of a red cell through the lung capillaries, such that alveolar and end-capillary PO2 are equal (and similarly for PCO2). In critically-ill patients in the ICU, this assumption is known to be reasonable.

Shunt

Shunt is defined as blood flowing from the right side to the left side of the heart without any alveolar gas contact, thus affording no O2 uptake or CO2 elimination. Shunt pathways include direct communications between the ventricles or atria of the heart, and blood passing through completely unventilated lung regions. Common causes include atelectasis, alveolar filling with fluid or exudate, and pneumothorax. Shunting is common and often substantial in the ICU. Uncommonly, dilated pulmonary vessels may carry blood that fails to (fully) oxygenate, even when adjacent alveoli are ventilated. This may happen in chronic liver disease [8]‌, and in infants/children suffering bronchopulmonary dysplasia.

Extrapulmonary factors

For a given metabolic rate ( V ˙ o 2 ), reduction in cardiac output will necessitate greater tissue O2 extraction to maintain adequate O2 supply. If cardiac output falls (in relation to V ˙ o 2 ), PO2 of the venous blood returning to the lungs also falls. Similarly, high cardiac output (in relation to V ˙ o 2 ) will result in an elevated venous PO2. As the ventilation/perfusion equation shows (eqn. 4), the boundary conditions (composition of pulmonary arterial blood and inspired gas) affect the values of PO2 and PCO2 independently of the V ˙ A / Q ˙ ratio. As venous PO2 falls, alveolar, and end-capillary PO2 will fall at a given V ˙ A / Q ˙ ratio, and vice versa. In this way, extrapulmonary factors (cardiac output and V ˙ o 2 especially) will affect arterial PO2, in addition to all the preceding potential causes of hypoxaemia. Reduction in blood [Hb] has the same effect as reduction in cardiac output—greater fractional extraction of O2 and lower venous PO2.

The message is that when changes in metabolism, cardiac function, and/or [Hb] occur, changes in arterial oxygenation are expected that have nothing to do with lung health. Recognizing this is paramount in providing the best therapeutic response. For example, thinking a sudden fall in arterial PO2 is due to new alveolar oedema or collapse when it is due to cardiac dysfunction may lead to poor clinical decisions.

In closing, two special cases of the role of extrapulmonary factors deserve mention—the first is dependence of shunt on total pulmonary blood flow. In patients with shunt, a change in cardiac output changes not just venous PO2 as discussed, but also the volumetric flow distribution between the shunt channels and the rest of the lung. The reason remains unclear, but may have to do with differing degrees of hypoxic pulmonary vasoconstriction in the shunt vessels compared to the rest of the lungs. Fig. 75.5 shows an extraordinary example of this effect [9]‌. The implication is clear—increasing blood flow through a damaged lung has the potential of worsening the shunt, which will oppose any beneficial effect of higher blood flow on the venous PO2.

Fig. 75.5 Change in shunt fraction and flow in one ICU patient [9]‌.

Fig. 75.5 Change in shunt fraction and flow in one ICU patient [9]‌.

Reproduced with permission of the authors. Lemaire F, Harf A, Teisseire BP, 'Oxygen exchange across the acutely injured lung'. In: Zapol WM, Falke KJ (eds), Acute Respiratory Failure, New York: Dekker; 1985, p. 521–52.

The second example occurs when high positive airway pressure are used. This impedes venous return to the heart and cardiac output falls. As a result, even if higher airway pressure successfully re-inflates collapsed alveoli and reduces shunt, the expected benefit in arterial PO2 may not occur, or may be lessened, because the reduced cardiac output lowers venous PO2.

References

1. Rahn H and Fenn WO. (1955). A graphical analysis of the respiratory gas exchange. Washington, DC: American Physiological Society.Find this resource:

2. Riley RL and Cournand A. (1951). Analysis of factors affecting partial pressures of oxygen and carbon dioxide in gas and blood of lungs: theory. Journal of Applied Physiology, 4, 77–101.Find this resource:

3. Riley RL, Cournand A, and Donald KW. (1951). Analysis of factors affecting partial pressures of oxygen and carbon dioxide in gas and blood of lungs: methods. Journal of Applied Physiology, 4, 102–20.Find this resource:

4. Riley RL and Houston CS. (1951) Composition of alveolar air and volume of pulmonary ventilation during long exposure to high altitude. Journal of Applied Physiology, 3, 526–34.Find this resource:

5. Young IH, Mazzone RW, and Wagner PD. (1980). Identification of functional lung unit in the dog by graded vascular embolization. Journal of Applied Physiology, 49(1), 132–41.Find this resource:

6. Wagner PD and West JB. (1972). Effects of diffusion impairment on O2 and CO2 time courses in pulmonary capillaries. Journal of Applied Physiology, 33(1), 62–71.Find this resource:

7. West JB and Dollery CT. (1964). Distribution of blood flow in isolated lung: relation to vascular and alveolar pressures. Journal of Applied Physiology, 19, 713–24.Find this resource:

8. Rodriguez-Roisin R and Krowka MJ. (2008). Hepatopulmonary syndrome—A liver-induced lung vascular disorder. New England Journal of Medicine, 358(22), 2378–87.Find this resource:

9. Lemaire F, Harf A, and Teisseire BP. (1985). Oxygen exchange across the acutely injured lung. In: Zapol WM, Falke KJ (eds) Acute respiratory failure, pp. 521–2. New York: Dekker.Find this resource: