Oxygen Delivery, Oxygen Consumption, and Cellular Hypoxia

An oxygen saturation of 98% does not establish that the tissues receive enough oxygen. The blood may contain too little hemoglobin, or too little blood may reach the organs. Even adequate total oxygen delivery can conceal poor perfusion in an individual tissue.

Oxygen transport must therefore be followed from the lungs to the blood, through the circulation, and into the cells. Each measurement describes a different part of that pathway.

4.1 Oxygen Tension, Saturation, and Content Describe Different Things

Arterial oxygen tension, arterial oxygen saturation, and arterial oxygen content are related but distinct measurements. Confusing them can lead to an incorrect assessment of oxygen transport.

Arterial oxygen tension, written PaO₂, is the partial pressure produced by oxygen dissolved in arterial blood. Oxygen moves from regions of higher oxygen tension toward regions of lower tension. PaO₂ also influences how readily oxygen binds to hemoglobin. It does not measure the total amount of oxygen in the blood.

Arterial oxygen saturation, written SaO₂, describes the proportion of hemoglobin’s oxygen-binding sites occupied by oxygen. The oxyhemoglobin dissociation curve relates saturation to oxygen tension. On its upper portion, most binding sites are already occupied. A large increase in PaO₂ therefore produces little further increase in saturation. Temperature, pH, and other factors alter hemoglobin’s oxygen affinity and can shift this curve.

Pulse oximetry estimates arterial saturation and displays SpO₂. It does not measure hemoglobin concentration, cardiac output, or oxygen content. Motion, poor perfusion, and other sources of measurement error can affect its interpretation. A blood gas result may also report a calculated saturation rather than one directly measured by co-oximetry. The measurement method matters when the clinical picture and the displayed value disagree.

Most arterial oxygen is carried on hemoglobin. A much smaller amount is physically dissolved in plasma under ordinary clinical conditions. A commonly used approximation is:

CaO₂ is arterial oxygen content, expressed as milliliters of oxygen in each deciliter of blood. One deciliter is 100 mL. Hb is hemoglobin concentration in g/dL. Enter SaO₂ as a fraction: 98% becomes 0.98. Enter PaO₂ in mm Hg.

The first term calculates oxygen carried on hemoglobin. The coefficient 1.34 represents the approximate milliliters of oxygen carried by each gram of fully saturated hemoglobin. The second term calculates dissolved oxygen. Its coefficient, 0.0031, gives the approximate milliliters dissolved per deciliter for each mm Hg of oxygen tension.

This equation assumes that abnormal hemoglobin species are not materially affecting oxygen carriage. Carbon monoxide exposure and methemoglobinemia require particular care because ordinary saturation readings may not adequately describe usable oxygen content.

Consider illustrative values of Hb 10 g/dL, SaO₂ 98%, and PaO₂ 100 mm Hg. First calculate oxygen carried on hemoglobin:

1.34 × 10 × 0.98 = 13.132 mL O₂/dL

Then calculate dissolved oxygen:

0.0031 × 100 = 0.31 mL O₂/dL

Adding them gives approximately 13.44 mL O₂/dL. Each 100 mL of arterial blood therefore carries about 13.44 mL of oxygen.

Now let Hb fall to 7 g/dL while saturation and tension remain unchanged. The calculation becomes:

(1.34 × 7 × 0.98) + 0.31 ≈ 9.50 mL O₂/dL

The hemoglobin that remains is still 98% saturated. There is simply less hemoglobin available to carry oxygen, so total content falls.

4.2 Oxygen Delivery: Content Multiplied by Flow

Systemic oxygen delivery is the amount of oxygen transported to the systemic circulation per unit time. It is written DO₂ and depends on both cardiac output and arterial oxygen content.

When cardiac output is in L/min and content is in mL O₂/dL, the factor of 10 converts liters to deciliters. The result is in mL O₂/min. Omitting this conversion creates a tenfold error.

Use the previous content of 13.44 mL O₂/dL. A cardiac output of 5 L/min moves 50 dL of blood each minute. Oxygen delivery is therefore:

5 × 13.44 × 10 = 672 mL O₂/min

If cardiac output falls to 3 L/min while content stays unchanged:

3 × 13.44 × 10 ≈ 403 mL O₂/min

The blood carries the same oxygen per unit volume, but less blood reaches the systemic circulation each minute. Delivery falls even though saturation and hemoglobin concentration have not changed.

The equation identifies three major determinants of delivery: cardiac output, hemoglobin concentration, and arterial oxygenation. Their importance depends on which is limiting. If saturation is very low, improving oxygenation fills more hemoglobin-binding sites and can substantially increase content. If hemoglobin is already almost fully saturated, raising PaO₂ mainly adds the much smaller dissolved component.

Cardiac output and oxygen delivery can also be indexed to body surface area. Cardiac index replaces cardiac output in the indexed calculation. The resulting oxygen delivery index has units of mL O₂/min/m². Indexed and unindexed values should not be compared as though they were the same measurement.

DO₂ describes total systemic transport. It does not establish how blood is divided between organs or whether oxygen reaches every cell within an organ. A useful global value must still be interpreted alongside the distribution of perfusion.

4.3 Oxygen Consumption and the Fick Relationship

Oxygen consumption is the rate at which the body takes up oxygen. It is written VO₂. In a stable system, the oxygen removed from blood across the systemic circulation equals the body’s oxygen consumption.

The Fick relationship expresses this conservation of oxygen:

CvO₂ is mixed venous oxygen content in mL O₂/dL. It represents blood returning from the systemic circulation after mixing in the right heart. The factor of 10 is again required when flow is expressed in L/min.

For an illustrative cardiac output of 5 L/min, let arterial content be 20 mL O₂/dL and mixed venous content be 15 mL O₂/dL. Subtract the returning content from the delivered content:

20 − 15 = 5 mL O₂ removed from each dL of blood

Then multiply by the volume of blood passing through each minute:

5 L/min × 5 mL O₂/dL × 10 = 250 mL O₂/min

The body is taking up 250 mL of oxygen each minute in this example. These assumed values are teaching examples, not treatment targets.

The relationship can be rearranged to calculate cardiac output from independently measured oxygen consumption and an arterial-to-mixed-venous content difference. If oxygen consumption is estimated rather than measured, error in that estimate becomes error in the calculated output. An assumed VO₂ should not be mistaken for an independent measurement of metabolic activity.

Oxygen consumption can also be assessed from respiratory gas exchange. This requires equipment and conditions that permit reliable measurement. Leaks, unstable conditions, and limitations of the measuring system can affect the result.

Calculated DO₂ and VO₂ share some of the same measurements. Both include cardiac output and arterial oxygen content. If cardiac output is overestimated, both calculated values rise even if the patient’s physiology has not changed. This creates mathematical coupling: two results appear related partly because they contain the same input. Phang and colleagues demonstrated this problem in ARDS. A rise in calculated consumption alongside calculated delivery does not, by itself, prove that tissues used more oxygen. 1

4.4 Oxygen Extraction and the Limits of Compensation

The oxygen extraction ratio describes the proportion of delivered oxygen taken up by the body.

In the preceding example, oxygen delivery is 5 × 20 × 10, or 1,000 mL O₂/min. Consumption is 250 mL O₂/min. Divide consumption by delivery:

250 ÷ 1,000 = 0.25 = 25%

The body takes up one quarter of the oxygen delivered in this example. Resting whole-body extraction is often near this value in healthy adults, but it varies between organs and with physiological conditions.

If less oxygen arrives each minute, tissues may maintain uptake by removing more from each unit of blood. Less oxygen then remains in the venous blood, so mixed venous oxygen content falls. This increased extraction can preserve consumption over a range of falling delivery.

Extraction cannot increase without limit. Once tissues cannot remove enough additional oxygen, a further fall in delivery lowers oxygen uptake. Consumption has become supply dependent: the oxygen arriving now limits the oxygen the tissues can use. The delivery at which this transition occurs is called the critical oxygen delivery threshold.

The threshold is a physiological concept rather than a single number that applies to every patient. It depends on metabolic demand, blood-flow distribution, diffusion, and the capacity to extract oxygen. A global calculation also combines tissues with different thresholds. One organ may become hypoxic while total consumption appears preserved.

Demand changes the interpretation. Fever, seizures, shivering, and increased respiratory muscle work can increase oxygen requirements. A delivery that was adequate before the increase in workload may become inadequate afterward.

Conversely, a reduction in measured consumption is not always reassuring. It may reflect lower demand after effective treatment or sedation. It may also reflect insufficient oxygen supply or impaired cellular use. The direction of change must be linked to the clinical circumstances.

4.5 Venous Oxygen Saturation: A Balance, Not a Verdict

Mixed venous oxygen saturation, written SvO₂, is measured in pulmonary arterial blood after systemic venous blood has mixed in the right heart. It describes the oxygen remaining on hemoglobin after passage through the tissues. How much remains depends on how much was delivered and how much was taken up. Arterial oxygenation, hemoglobin concentration, cardiac output, and oxygen consumption therefore all affect SvO₂.

A low SvO₂ indicates that returning blood contains relatively little oxygen on hemoglobin. Possible explanations include reduced delivery, increased demand, or both. The value does not identify which determinant changed.

A high SvO₂ has several possible explanations. Delivery may be high relative to demand. Demand may have fallen. Some blood may have passed through poorly extracting regions, or cellular extraction may be impaired. A high value therefore does not establish that all tissues are adequately oxygenated.

Central venous oxygen saturation, written ScvO₂, is usually sampled from the superior vena cava or near its junction with the right atrium. It contains a different mixture of venous drainage from pulmonary arterial blood. In particular, blood from the lower body and coronary circulation is not represented in the same proportions. ScvO₂ and SvO₂ can therefore differ even when both measurements are technically correct.

Chawla and colleagues compared paired central and mixed venous samples in critically ill patients. The measurements were not equivalent. Substituting central venous values into calculations that required mixed venous blood produced substantial errors. A fixed correction cannot reliably make the measurements interchangeable. 2

High values can also accompany poor outcomes. Textoris and colleagues retrospectively examined ScvO₂ measurements obtained within the first 72 hours of septic shock. Higher maximum values were associated with increased hospital mortality. The study does not show that high saturation caused harm or that lowering it would help. It shows why a high ScvO₂ alone cannot establish successful resuscitation. 3

Trends can add information when sampling location and measurement conditions remain comparable. Nevertheless, a change in venous saturation still requires an explanation based on the wider clinical assessment.

4.6 Cellular Hypoxia Can Occur Despite Adequate Global Delivery

Oxygen must move from perfused capillaries into the surrounding tissue. Delivery can fail at the level of the whole circulation, within an organ, or across the path between the capillary and the cell.

A low cardiac output reduces the total blood available to distribute. A regional arterial obstruction can interrupt delivery to one territory despite an adequate total output. Uneven capillary perfusion can leave some cells poorly supplied while nearby vessels carry substantial flow.

De Backer and colleagues demonstrated abnormal small-vessel perfusion in patients with sepsis. The abnormalities were more pronounced in patients with worse outcomes. This human study supports the importance of microvascular distribution, while its observational design cannot establish that correcting a particular microcirculatory measurement improves survival. 4

Oxygen must also diffuse from capillary blood to the cell. Its movement depends on the local oxygen-tension difference and the distance it must cross. Edema can increase the separation between a perfused capillary and nearby cells. Oxygen may then reach those cells less readily, although the effect depends on the tissue and its microvascular arrangement.

Hemoglobin must release oxygen as well as carry it. Higher oxygen affinity means that hemoglobin holds oxygen more tightly at a given oxygen tension. This can reduce unloading in tissue under otherwise comparable conditions. Arterial saturation describes how well hemoglobin has loaded oxygen; it does not show how much oxygen subsequently leaves capillary blood.

Finally, cells must use the oxygen that arrives. As discussed in Chapter 3, altered mitochondrial function can impair oxidative energy production. This is different from insufficient oxygen availability, although both can occur together. Routine systemic measurements cannot reliably separate them on their own.

These mechanisms explain the limits of global reassurance. A satisfactory DO₂, SvO₂, or lactate does not independently exclude a regional perfusion emergency. Focal clinical findings still require direct investigation.

4.7 Lactate and Oxygen Debt

When oxygen supply cannot meet demand, oxidative energy production may fall and lactate may rise. In this setting, lactate can be an important signal of a delivery–demand mismatch. A rising concentration during clinical deterioration deserves prompt attention.

Oxygen debt is the shortfall that accumulates when oxygen uptake remains below the amount required over time. Both the size of the mismatch and its duration determine the accumulated shortfall. A lactate concentration measures neither of these directly, so it cannot quantify oxygen debt.

Lactate can rise through accelerated glycolysis even when oxygen is present. James and colleagues demonstrated this response to epinephrine in oxygenated rat skeletal muscle. Reduced lactate removal and genuine hypoperfusion can coexist with increased production. 5

A normal lactate also has limits. A localized ischemic process may not generate a large enough systemic signal to raise the blood concentration. Early disease and the balance between production and removal can further affect the result.

Repeated measurements help describe the direction of change, but they must remain connected to the patient. A falling lactate with improved perfusion is encouraging. Persistent lactate after the circulation improves should prompt reassessment of both perfusion and alternative metabolic explanations. It does not automatically establish the need for additional fluid or an inotrope.

Lactate is therefore most useful as part of a physiological assessment. Its value comes from the information it adds to examination, hemodynamics, and organ function rather than from treating it as a direct oxygen meter.

4.8 Improving the Balance Between Delivery and Demand

An intervention should address the determinant that is limiting oxygen availability. Increasing a variable that is already adequate may produce little useful benefit.

When arterial oxygenation is inadequate, respiratory treatment can improve oxygen loading. The underlying cause may require oxygen supplementation, ventilatory support, recruitment of functioning lung, or treatment of the pulmonary disease. The appropriate intervention depends on the mechanism and its effects on the circulation.

When cardiac output is inadequate, identify what limits ejection. Additional filling can increase output if the ventricles remain preload responsive. The same fluid can also raise filling pressures and worsen edema. If poor contraction, excessive afterload, an obstruction, or an arrhythmia limits flow, adding fluid may not correct the problem. A low calculated delivery therefore identifies a transport concern, not the treatment needed to correct it.

Increasing hemoglobin can increase oxygen content. The clinical value depends on the severity and cause of anemia, the adequacy of blood flow, and the consequences of transfusion. An increase in calculated DO₂ is not sufficient evidence that a transfusion will improve organ function.

Demand also requires attention. Treating seizures, shivering, pain, and excessive respiratory effort can reduce an unsustainable workload. The intervention itself may have other effects. Sedation can reduce activity but also lower arterial pressure or delay assessment. Positive-pressure ventilation can unload respiratory muscles while altering venous return and right ventricular function.

The intended response should be defined before treatment. Evidence of benefit may include improved mentation, peripheral perfusion, cardiac output, or organ function. The timing of reassessment should reflect how quickly the relevant variable can change. Kidney recovery, for example, may lag behind improvement in circulation.

4.9 What Treatment Trials Add to the Physiology

Physiology explains how an intervention can change oxygen transport. Clinical trials test whether a particular strategy improves outcomes in a defined population.

Hayes and colleagues randomized 100 critically ill patients who had not achieved predetermined oxygen-transport goals with volume expansion alone. A dobutamine-based strategy increased cardiac index and delivery but did not increase oxygen consumption. Hospital mortality was higher in the intervention group. The study argues against aggressively pursuing supranormal transport targets in that setting. It does not establish that dobutamine is inappropriate when a patient has a specific indication for inotropic support. 6

TRISS compared transfusion thresholds of 7 and 9 g/dL in septic shock. Ninety-day mortality did not differ significantly, and the lower-threshold group received fewer transfusions. The higher threshold therefore did not demonstrate a survival advantage in the studied population. Acute coronary syndrome and life-threatening bleeding were exclusions. The protocol could also be suspended if life-threatening bleeding or ischemia developed. These settings require separate clinical judgment. 7

Oxygen therapy also requires outcome evidence. The UK-ROX trial randomized 16,500 mechanically ventilated adults receiving supplemental oxygen. A conservative strategy targeting SpO₂ of 90%, within a range of 88%–92%, did not significantly reduce 90-day mortality compared with usual care. This trial does not justify uncontrolled hyperoxia or establish one target for every diagnosis. It demonstrates that reducing oxygen exposure through this particular strategy did not produce the anticipated overall mortality benefit. 8

These studies address different interventions and populations. Their shared lesson is that improving a transport variable, or reducing an exposure, must be connected to an outcome that matters to the patient. The treatment strategy, timing, and associated harms all influence that connection.

4.10 Applying Oxygen Transport at the Bedside

Consider an illustrative patient with septic shock. Hemoglobin is 8 g/dL. SaO₂ is 98%, PaO₂ is 100 mm Hg, and measured cardiac output is 3 L/min. The patient has cool extremities, prolonged capillary refill, and rising lactate.

First calculate the oxygen carried in each deciliter of blood:

CaO₂ = (1.34 × 8 × 0.98) + (0.0031 × 100) ≈ 10.82 mL O₂/dL

Next multiply by cardiac output and convert liters to deciliters. Using the full content calculation:

DO₂ = 3 × [(1.34 × 8 × 0.98) + (0.0031 × 100)] × 10 ≈ 324 mL O₂/min

The saturation is satisfactory, but low flow and reduced hemoglobin limit total transport. Together with the cool extremities, delayed capillary refill, and rising lactate, this raises concern about inadequate delivery. The calculation does not establish the exact oxygen requirement or identify one treatment.

Now consider an arithmetic comparison in which SaO₂ rises from 98% to 100% and PaO₂ rises from 100 to 300 mm Hg. With Hb still 8 g/dL:

CaO₂ = (1.34 × 8 × 1.00) + (0.0031 × 300) = 11.65 mL O₂/dL

At the same cardiac output of 3 L/min:

DO₂ = 3 × 11.65 × 10 ≈ 350 mL O₂/min

The higher PaO₂ adds some oxygen, but it leaves the low-flow problem unchanged. These assumed values illustrate the calculation; they are not a recommendation to raise PaO₂ to 300 mm Hg.

Return to the original oxygen content and instead increase cardiac output from 3 to 4.5 L/min in the calculation:

DO₂ ≈ 4.5 × 10.82 × 10 ≈ 487 mL O₂/min

More blood passes through each minute, carrying more oxygen in total. This illustrates the importance of flow. It does not establish whether fluid, an inotrope, or another intervention can safely produce that increase in this patient.

The next assessment should establish why output is low. Echocardiography, the rhythm, ventricular loading conditions, and an appropriate assessment of fluid responsiveness can help distinguish mechanisms. Congestion and the risk of fluid accumulation must also be considered. The hemoglobin result should be interpreted alongside bleeding, ischemia, and the broader transfusion context.

Suppose a central venous sample gives an ScvO₂ of 80%. Much of the oxygen remains on hemoglobin in that sample, but the value does not explain why. Reduced extraction or uneven distribution could contribute. Check sampling and measurement quality as well. The value does not cancel the evidence of poor perfusion and cannot replace a mixed venous measurement in a global oxygen-consumption calculation.

The case illustrates the purpose of oxygen-transport calculations. They make the relevant determinants visible. The physician must then connect those determinants to the patient’s mechanism, choose a treatment with a defined purpose, and assess the resulting change in perfusion and organ function.

References

1. Phang PT, Cunningham KF, Ronco JJ, Wiggs BR, Russell JA. Mathematical coupling explains dependence of oxygen consumption on oxygen delivery in ARDS. American Journal of Respiratory and Critical Care Medicine. 1994;150(2):318–323. doi:10.1164/ajrccm.150.2.8049809.

2. Chawla LS, Zia H, Gutierrez G, Katz NM, Seneff MG, Shah M. Lack of equivalence between central and mixed venous oxygen saturation. Chest. 2004;126(6):1891–1896. doi:10.1378/chest.126.6.1891.

3. Textoris J, Fouché L, Wiramus S, et al. High central venous oxygen saturation in the latter stages of septic shock is associated with increased mortality. Critical Care. 2011;15:R176. doi:10.1186/cc10325.

4. De Backer D, Creteur J, Preiser JC, Dubois MJ, Vincent JL. Microvascular blood flow is altered in patients with sepsis. American Journal of Respiratory and Critical Care Medicine. 2002;166(1):98–104. doi:10.1164/rccm.200109-016OC.

5. James JH, Wagner KR, King JK, et al. Stimulation of both aerobic glycolysis and Na+-K+-ATPase activity in skeletal muscle by epinephrine or amylin. Am J Physiol. 1999;277:E176–E186. Original experimental study.

6. Hayes MA, Timmins AC, Yau EHS, Palazzo M, Hinds CJ, Watson D. Elevation of systemic oxygen delivery in the treatment of critically ill patients. New England Journal of Medicine. 1994;330(24):1717–1722. doi:10.1056/NEJM199406163302404.

7. Holst LB, Haase N, Wetterslev J, et al. Lower versus higher hemoglobin threshold for transfusion in septic shock. New England Journal of Medicine. 2014;371(15):1381–1391. doi:10.1056/NEJMoa1406617.

8. Martin DS, Gould DW, Shahid T, et al. Conservative oxygen therapy in mechanically ventilated critically ill adult patients: the UK-ROX randomized clinical trial. JAMA. 2025;334(5):398–408. doi:10.1001/jama.2025.9663.