The alveolar-arterial (A-a) oxygen gradient, often abbreviated as A-a gradient or A-aDO2, is the difference between the partial pressure of oxygen in the alveoli (PAO2) and the partial pressure of oxygen in arterial blood (PaO2). In a healthy lung with perfect gas exchange, these values should be nearly identical. In reality, even in healthy individuals, a small A-a gradient exists due to normal physiological shunting (bronchial and thebesian veins draining deoxygenated blood directly into the arterial circulation) and minor ventilation-perfusion (V/Q) mismatching. The alveolar PO2 (PAO2) is calculated from the alveolar gas equation using the fraction of inspired oxygen (FiO2), atmospheric pressure (Patm, normally 760 mmHg at sea level), water vapour pressure (PH2O = 47 mmHg at 37°C), and the arterial CO2 tension (PaCO2) corrected by the respiratory quotient (RQ = 0.8). The PaO2 is measured directly from an arterial blood gas (ABG) sample. A normal A-a gradient in young adults is less than 10 mmHg on room air. It increases with age at approximately 1 mmHg per decade (or 1 mmHg per 4 years of age), reflecting age-related V/Q mismatching. An elevated A-a gradient indicates an intrinsic pulmonary cause of hypoxaemia: V/Q mismatch (as in pulmonary embolism, pneumonia, asthma), diffusion impairment (interstitial lung disease), or intracardiac/intrapulmonary shunting. Importantly, the A-a gradient is normal in hypoventilation (e.g., opioid overdose, neuromuscular disease) and at altitude — conditions where PaO2 is low but the alveoli are similarly hypoxic. This makes the A-a gradient an essential diagnostic tool to distinguish between pulmonary and extra-pulmonary causes of hypoxaemia.
PAO2 = FiO2 × (Patm − PH2O) − PaCO2 / RQ
= FiO2 × (760 − 47) − PaCO2 / 0.8
= FiO2 × 713 − PaCO2 / 0.8
A-a Gradient = PAO2 − PaO2
Normal: <10 mmHg (young adult on room air)
Age-adjusted normal: Age / 4 + 4 (mmHg)- 1Obtain an arterial blood gas (ABG) sample, noting PaO2, PaCO2, pH, and the FiO2 being delivered at time of sampling.
- 2Note the atmospheric pressure — at sea level Patm = 760 mmHg; reduce for altitude (e.g., 640 mmHg at 2,000 m altitude).
- 3Calculate PAO2 using the alveolar gas equation: PAO2 = FiO2 × (760 − 47) − PaCO2 / 0.8.
- 4Subtract PaO2 (from ABG) from PAO2 to obtain the A-a gradient.
- 5Determine the age-adjusted normal A-a gradient: Age/4 + 4 mmHg; compare with calculated value.
- 6Interpret: A-a gradient within normal = likely extra-pulmonary cause of hypoxaemia (hypoventilation, low FiO2). Elevated A-a gradient = pulmonary cause (V/Q mismatch, shunt, diffusion).
- 7Combine with clinical context, chest X-ray, and CTPA/V/Q scan as appropriate to identify the specific pulmonary pathology.
A-a gradient 56.7 >> normal 17.75 — pulmonary cause confirmed; PE workup warranted
The markedly elevated A-a gradient with hypoxaemia and hypocarbia (from hyperventilation) is a classic PE presentation. The normal A-a gradient would be ~18 mmHg. The gap of nearly 57 mmHg confirms a significant pulmonary cause — V/Q mismatch from pulmonary emboli preventing adequate gas exchange.
A-a gradient 9.7 — normal; hypoventilation from opioid overdose, no primary lung disease
The elevated PaCO2 confirms hypoventilation. However, the A-a gradient is normal — the alveoli and arterial blood are both hypoxic, but gas exchange at the alveolar level is intact. This pattern indicates extra-pulmonary hypoventilation (opioid, sedative, neuromuscular disease) rather than primary lung disease.
Elevated A-a gradient on supplemental oxygen — significant V/Q mismatch or shunt
On supplemental oxygen, the A-a gradient is normally larger but should not be this high. A PaO2 of 85 mmHg on 40% FiO2 (corresponding P/F ratio = 212) with an A-a gradient of 148 mmHg confirms severe impairment of gas exchange consistent with ARDS or severe pneumonia.
Normal — within expected range for age
A-a gradient of 1.7 mmHg in a healthy young adult is well within the normal range. It represents the minimal physiological shunting that exists even in normal lungs. This baseline helps understand how significantly the gradient rises in pathological states.
Emergency department assessment of dyspnoeic patients to quickly determine whether hypoxaemia is pulmonary or extra-pulmonary in origin, enabling practitioners to make well-informed quantitative decisions based on validated computational methods and industry-standard approaches
ICU monitoring of ARDS progression and response to ventilator management strategies including PEEP optimisation and prone positioning, helping analysts produce accurate results that support strategic planning, resource allocation, and performance benchmarking across organizations
Pulmonary embolism clinical decision pathways where a normal A-a gradient raises the pre-test probability but cannot rule out PE, allowing professionals to quantify outcomes systematically and compare scenarios using reliable mathematical frameworks and established formulas
Anaesthetic assessment of ventilation-perfusion matching before and after intubation in high-risk surgical patients, supporting data-driven evaluation processes where numerical precision is essential for compliance, reporting, and optimization objectives, necessitating robust computational methods that deliver consistent and verifiable results suitable for reporting, auditing, and long-term trend analysis in professional environments
Acclimatisation assessment at altitude research stations to distinguish physiological versus pathological oxygen levels, which requires precise quantitative analysis to support evidence-based decisions, strategic resource allocation, and performance optimization across diverse organizational contexts and professional disciplines
Altitude and Reduced Atmospheric Pressure
At altitude, Patm is lower (e.g., 480 mmHg at 3,500 m), which reduces PAO2. The A-a gradient remains normal in healthy altitude-exposed individuals because gas exchange across the alveolar membrane is intact. Always use the local atmospheric pressure when calculating the A-a gradient in patients at high altitude or in pressurised aircraft (Patm approximately 565 mmHg).
Intracardiac Right-to-Left Shunt
In conditions such as atrial septal defect (ASD), ventricular septal defect (VSD), or patent foramen ovale (PFO) with right-to-left flow, deoxygenated blood enters the arterial circulation without passing through the lungs. This raises the A-a gradient significantly. Crucially, shunt hypoxaemia does not respond well to supplemental oxygen — a hallmark that distinguishes shunt from V/Q mismatch.
Mechanical Ventilation
In ventilated patients, the FiO2 is directly set on the ventilator and is more precisely known than in spontaneously breathing patients on facemasks. The A-a gradient can be calculated more accurately in ventilated patients, and serial measurements help monitor the response to PEEP adjustments, prone positioning, and recruitment manoeuvres in ARDS management.
Carbon Monoxide Poisoning
CO poisoning causes severe tissue hypoxia despite normal (or near-normal) PaO2 and a normal A-a gradient. PaO2 is the dissolved oxygen tension in plasma and is unaffected by CO binding to haemoglobin. Co-oximetry (measuring COHb) is essential in suspected CO poisoning; standard pulse oximetry and ABG O2 measurements are misleading.
| A-a Gradient | Interpretation | Common Causes |
|---|---|---|
| <10 mmHg (young) / <Age/4+4 | Normal | No pulmonary gas exchange impairment |
| 10–30 mmHg | Mildly elevated | Early V/Q mismatch, mild heart failure, early pneumonia |
| 30–60 mmHg | Moderately elevated | Pneumonia, PE, COPD exacerbation, moderate pulmonary oedema |
| >60 mmHg | Severely elevated | ARDS, severe pneumonia, large shunt, severe pulmonary oedema |
What is the respiratory quotient (RQ) and why is 0.8 used?
The respiratory quotient (RQ) is the ratio of CO2 produced to O2 consumed in cellular metabolism. It depends on the metabolic substrate being used: 1.0 for pure carbohydrates, 0.7 for fats, and 0.8 for a mixed diet. 0.8 is used as a standard approximation for the average mixed-substrate diet in clinical calculations. Using the correct RQ is important only when dietary intake is very specific (e.g., parenteral nutrition with high carbohydrate load).
Why does the A-a gradient increase with age?
Ageing causes progressive loss of alveolar elasticity, airway closure at higher lung volumes, and reduced ventilation of dependent lung zones. This increases physiological V/Q mismatching. Each decade adds approximately 2.5–4 mmHg to the normal A-a gradient. Using the age-adjusted formula (Age/4 + 4) prevents falsely flagging age-related physiological changes as pathological.
Can the A-a gradient be elevated without clinical hypoxaemia?
Yes. In early or mild lung disease, the A-a gradient may be elevated before overt hypoxaemia develops, particularly if the patient is hyperventilating and maintaining a normal PaO2 through compensatory increased minute ventilation. An elevated A-a gradient with normal PaO2 is an early sensitive sign of pulmonary disease, such as early PE or mild interstitial lung disease.
How is the A-a gradient affected by breathing supplemental oxygen?
On supplemental oxygen, the A-a gradient is normally larger than on room air because a higher PAO2 is expected. The gradient can be 100+ mmHg even in healthy individuals at high FiO2. Therefore, absolute A-a gradient values cannot be directly compared across different FiO2 levels. The P/F ratio (PaO2/FiO2) is preferred for comparing gas exchange efficiency across oxygen concentrations.
What causes an elevated A-a gradient?
The three main mechanisms are: (1) V/Q mismatch — the most common cause, seen in PE, pneumonia, asthma, COPD, heart failure; (2) intrapulmonary or intracardiac shunt — where blood bypasses ventilated alveoli entirely, seen in ARDS, large pneumonia consolidation, atrial septal defect; (3) diffusion impairment — oxygen cannot cross the alveolar membrane efficiently, as in interstitial lung disease and pulmonary fibrosis.
Is an elevated A-a gradient specific for pulmonary embolism?
No. An elevated A-a gradient is highly sensitive but not specific for PE. It is elevated in any condition causing V/Q mismatch, shunt, or diffusion abnormality. Importantly, approximately 10–15% of confirmed PEs can present with a normal A-a gradient, particularly in young patients with small emboli who can compensate with hyperventilation. The A-a gradient alone cannot diagnose or exclude PE.
What is the relationship between A-a gradient and the P/F ratio?
Both assess gas exchange efficiency. The A-a gradient requires a calculated PAO2, while the P/F ratio (PaO2/FiO2) is simpler and used directly from ABG values without calculating PAO2. The P/F ratio is used in ARDS staging (Berlin criteria) and SOFA score. The A-a gradient provides more mechanistic information but is less practical for titrating oxygen therapy at varying FiO2 levels.
What is the A-a gradient at altitude?
At altitude, the lower atmospheric pressure reduces PAO2 even at the same FiO2. Acclimatised individuals hyperventilate to reduce PaCO2, partially compensating. The A-a gradient is typically normal at altitude (confirming gas exchange is intact), while PaO2 is reduced because of the lower atmospheric oxygen tension. This is why climbers at altitude are hypoxic without inherent lung disease.
Pro Tip
When evaluating hypoxaemia, always determine whether the A-a gradient is normal or elevated first. A normal A-a gradient directs you to extra-pulmonary causes (hypoventilation, low FiO2, altitude) and an elevated gradient confirms an intrinsic pulmonary problem. This single step organises the entire differential diagnosis of hypoxaemia in a logical, efficient way.
Did you know?
The alveolar gas equation was first formally described in the 1940s. The lungs' ability to perform gas exchange is extraordinary — with approximately 480 million alveoli providing a total surface area of 130 square metres (roughly the size of a singles tennis court), all folded into a volume barely larger than a basketball. The A-a gradient reflects how efficiently this vast surface is being used.
References
- ›West JB. Pulmonary Pathophysiology: The Essentials. 9th edition. Lippincott Williams & Wilkins. 2016.
- ›Petersson J, Glenny RW. Gas exchange and ventilation-perfusion relationships in the lung. Eur Respir J. 2014;44(4):1023-1041.
- ›Harris EA et al. The normal alveolar-arterial oxygen-tension gradient in man. Clin Sci Mol Med. 1974;46(1):89-104.
- ›Mellemgaard K. The alveolar-arterial oxygen difference: its size and components in normal man. Acta Physiol Scand. 1966;67(1):10-20.