A Physiological Model of Gas Exchange: How Lungs Move Oxygen and Carbon Dioxide

A physiological model of gas exchange describes oxygen and carbon dioxide moving between alveolar air, pulmonary capillary blood, and tissue cells purely by passive diffusion down partial-pressure gradients — no active pumping is involved at the membrane. In the lungs, the alveolar–arterial gradient for oxygen is roughly 60 mmHg at rest, driving O₂ across a membrane averaging 0.6 microns thick in about 0.25 seconds. In tissues, the same physics runs in reverse. Four steps govern the whole system: ventilation (moving gas in and out), diffusion (crossing the membrane), perfusion (blood flow carrying gas away), and chemical binding (hemoglobin capture and release). Most textbooks list these steps; the numbers below explain why each threshold — a shunt fraction above 30%, a diffusion time under 0.25 seconds — predicts failure before symptoms appear.

The Four-Step Mechanism: Ventilation, Diffusion, Perfusion, Binding

Ventilation is the only step that consumes energy. The diaphragm and intercostals generate negative intrathoracic pressure, pulling about 500 mL of air into the lung per quiet breath at 12–20 breaths per minute. That gives an alveolar ventilation of roughly 4–5 L/min — the figure doctors substitute into the alveolar gas equation, not the 6 L/min of minute ventilation, because 150 mL of every 500 mL breath never reaches an exchanging alveolus.

Diffusion follows Fick's law: flux is proportional to surface area and the partial-pressure gradient, and inversely proportional to membrane thickness and gas molecular weight. The lung's effective surface area is about 50–70 square meters — roughly the size of a tennis court folded into the chest — and the membrane thickness is 0.2 to 0.6 microns at its thinnest points. Carbon dioxide diffuses about 20 times faster than oxygen because its solubility in the membrane is higher, which is why CO₂ retention is a late sign in most lung disease.

Perfusion delivers roughly 5 L/min of mixed venous blood to the pulmonary capillaries. During its transpulmonary transit of about 0.75 seconds at rest, the capillary reaches 75% oxygen saturation in the first 0.25 seconds and is essentially fully equilibrated by 0.75 seconds. This reserve is the reason exercise, which shortens transit time, does not normally drop arterial oxygen.

Binding is the chemical amplification layer. Each hemoglobin molecule carries four oxygen molecules cooperatively, which produces the sigmoid dissociation curve and the P50 of 27 mmHg — the partial pressure at which hemoglobin is half-saturated. In plain terms: diffusion gets oxygen to the blood, but hemoglobin is what keeps it there long enough to deliver it to tissue.

Key Numbers Defined and Reconciled

These thresholds appear throughout clinical decision-making and must be read together, not separately. A normal shunt fraction is under 5%, but the same physiology becomes a clinical emergency when shunt exceeds 30% — the figure cited in critical care guidelines for hypoxemia uncorrected by supplemental oxygen. The alveolar–arterial oxygen gradient, normally under 15–20 mmHg in a young adult, widens above 30 mmHg with age, and any gradient above that range suggests diffusion or shunt failure rather than simple hypoventilation. The diffusing capacity for carbon monoxide (DLCO), the standard measure of membrane function, is normally above 75% of predicted; values under 60% predict oxygen desaturation during exercise. Hemoglobin normal range is 13.5–17.5 g/dL in men and 12–15 g/dL in women, and the oxygen delivery equation (DO₂) — cardiac output times arterial oxygen content — needs about 1,000 mL/min at rest. These figures interlock: a 30% shunt with a normal cardiac output still delivers less than 75% of baseline oxygen content, regardless of what the pulse oximeter reads.

The Oxygen–Hemoglobin Dissociation Curve: Where Diffusion Meets Delivery

The dissociation curve is the model's most misunderstood component. Hemoglobin binding is cooperative: the first O₂ molecule binds with difficulty, the second and third bind more readily, and the fourth binds easily. This produces the steep sigmoid shape and the flat upper plateau that keeps arterial saturation near 97–98% even when alveolar oxygen falls by 20 mmHg.

Three variables shift the curve and change delivery without changing saturation. A right shift moves P50 up toward 30 mmHg, releasing oxygen more readily to tissue; a left shift moves P50 down and holds oxygen too tightly. Temperature, pH, and 2,3-DPG control the direction. Fever or acidosis shifts right; hypothermia, alkalosis, and stored blood low in 2,3-DPG shift left. A hemoglobin frozen at 98% saturation during a left shift can underdeliver to tissue even while the pulse oximeter reports a reassuring number — the mechanism behind the counterintuitive finding that some patients with normal saturation remain hypoxic.

The Bohr effect and the Haldane effect run in parallel. The Bohr effect explains oxygen unloading in active tissue: CO₂ production lowers pH, shifting the curve right. The Haldane effect explains why deoxygenated blood carries more CO₂: oxygenated hemoglobin binds CO₂ less tightly, which is why the lung can unload CO₂ efficiently at the alveolus. Neither effect appears on any oximeter.

A Timeline of Gas Transit Through the Body

Gas exchange is a time-resolved process, not a static state. The following timeline traces one molecule of oxygen from the airway to the mitochondrion and one molecule of carbon dioxide back out.

Shunt, Dead Space, and Ventilation–Perfusion Mismatch

The model's most clinically useful predictions come from applying the ideal gas equation to non-ideal lungs. Two failure modes dominate. Shunt is blood that reaches the arterial system without seeing ventilated alveoli — in effect, right-to-left diversion. A shunt below 5% is normal because of the small bronchial and thebesian circulations; above 30%, supplemental oxygen barely raises arterial saturation because the shunted blood never contacts an alveolus holding extra oxygen.

Dead space is the reverse: ventilated space that receives no blood flow. Anatomical dead space is about 150 mL; physiological dead space can exceed 300 mL in pulmonary embolism or severe emphysema, and the fraction of tidal volume wasted as dead space rises above 30% in those conditions. At a tidal volume of 500 mL, a 30% dead-space fraction leaves only 350 mL for gas exchange per breath, requiring higher minute ventilation to maintain the same alveolar gas tension.

The non-obvious correction to the standard model: many search results describe gas exchange as a matter of "pressure gradients alone," which is incomplete. The model requires all four steps to be simultaneously functional. A patient with a normal membrane but a 50% shunt can have a normal DLCO and a near-normal diffusion step yet still fail arterial oxygenation. Membrane integrity does not equal gas exchange adequacy.

Distinguishing Established Physiology From Active Hypotheses

Most of this model is established science with decades of experimental replication. The alveolar gas equation, the Fick principle, and the sigmoid shape of the dissociation curve are textbook facts supported by direct measurement. The four-step framing appears in physiology curricula from major medical schools and is used to interpret arterial blood gases.

Three areas remain contested or actively researched. First, the exact upper limit of capillary transit time reserve during maximal exercise is debated: some studies report full equilibration down to 0.25 seconds in elite athletes, while others measure a diffusion limitation at maximal oxygen uptake. Second, the relative contribution of "microvascular recruitment" versus "capillary distension" to increasing the diffusion surface during exercise remains an open question. Third, the mechanism by which chronic hypoxia remodels the pulmonary vasculature, raising pulmonary artery pressure, is established in effect but the signaling cascade is still being mapped. A responsible model labels these as areas where certainty is incomplete — the rest of the framework is not.

Where the Model Predicts Failure: Edge Cases That Defy the Oximeter

The highest-value use of any physiological model is to predict what the measurement will miss. The pulse oximeter reads saturation, not content, and not delivery. Four edge cases break the oximeter's reassurance.

Carbon monoxide poisoning produces a falsely high saturation because carboxyhemoglobin absorbs light differently from oxyhemoglobin and is read as oxygenated. A patient with 40% carboxyhemoglobin can show 98% saturation on a standard pulse oximeter.

Anemia with normal saturation is the second case. At a hemoglobin of 7 g/dL, saturation can be 98% while oxygen content falls by 50%, because content equals 1.34 × Hb × saturation plus dissolved oxygen.

A left-shifted dissociation curve is the third: hemoglobin holding oxygen too tightly produces normal saturation with impaired tissue release. The fourth is high-altitude physiology, where hypoxic pulmonary vasoconstriction raises pulmonary artery pressure and redistributes blood flow to better-ventilated regions — a protective mechanism that, in the normal lung, already runs at 100% efficiency only because of its regional variability.

The practical action each edge case implies is a specific measurement, not a general caution:

• CO poisoning: measure carboxyhemoglobin with co-oximetry, not pulse oximetry
• Severe anemia: calculate oxygen content and delivery, not just saturation
• Suspected left shift: interpret saturation alongside lactate and mixed venous saturation
• High altitude: measure arterial blood gas if available, and expect a widened A–a gradient

What This Model Means for Everyday Health Decisions

For a healthy adult, the model's most useful output is the exercise reserve. Because capillary equilibration completes in 0.25 seconds out of a 0.75-second transit, the lung has a threefold time reserve before diffusion limits exercise. This explains why healthy people do not desaturate during moderate exercise, and why any oxygen desaturation during exercise should prompt evaluation rather than reassurance.

Three concrete actions follow directly from the model. Keep the membrane healthy by avoiding chronic irritants — the model predicts that any process thickening the 0.6-micron membrane widens the A–a gradient before symptoms. Maintain hemoglobin in the normal range through iron and B12 status, since content delivery depends on it. And treat breathlessness that worsens during exertion as a signal to measure, not dismiss: the model's thresholds (A–a gradient above 30, DLCO below 60%) are the point at which the physiology changes from compensated to failing.

For clinicians, the model's central action is to read every measurement alongside its physiological partner. Saturation is meaningless without hemoglobin. Alveolar ventilation is meaningless without dead-space fraction. DLCO is meaningless without shunt fraction. The system only fails when one step is compromised and the compensation can no longer hide it — which is exactly the point where the numbers on this page become the diagnosis.