Oxygen Transport: Oxygen-Hemoglobin Dissociation Curve & Factors Affecting Affinity (Bohr Effect, 2,3-DPG, PH, Temperature)

Foundations and mechanisms

Biophysical architecture of haemoglobin

Adult haemoglobin A (HbA) is an α₂β₂ tetramer; α chains contain 141 amino acids and β chains 146 amino acids. Each globin encloses one haem moiety: protoporphyrin IX with central ferrous iron, Fe²⁺, capable of reversibly binding one O₂ molecule. Thus each Hb molecule carries up to four O₂ molecules. Iron is coordinated to four pyrrole nitrogens, the proximal histidine F8, and, when oxygenated, O₂ at the sixth coordination site. Oxidation to Fe³⁺ forms methaemoglobin, which cannot bind O₂ effectively and shifts remaining ferrous haem sites functionally leftwards.

The central physiological property of Hb is that it is not a simple carrier but an allosteric molecular machine. It exists predominantly in two conformational ensembles: the low-affinity T state and high-affinity R state. Deoxyhaemoglobin is stabilised by salt bridges and a wide central cavity, particularly between β chains. Oxygenation narrows the central cavity and disrupts ionic interactions, reducing affinity for 2,3-diphosphoglycerate and protons.

Perutz mechanism and cooperativity

The Perutz model explains positive cooperativity structurally. In deoxyhaemoglobin, Fe²⁺ lies approximately 0.4 Å out of the porphyrin plane. O₂ binding pulls Fe²⁺ into the plane, dragging the proximal histidine and F helix. This movement is transmitted across the α₁β₂ and α₂β₁ interfaces, producing an approximately 15° rotation between αβ dimers. The affinity of unoccupied haem sites therefore increases after each O₂ molecule binds. This cooperative transition is responsible for the sigmoid oxygen-haemoglobin dissociation curve (OHDC), permitting near-complete loading in the lung while preserving substantial unloading over the tissue PO₂ range.

Hill representation and thermodynamic interpretation

Let fractional saturation be Y, equivalent clinically to S_O2 when expressed as a fraction. A simplified Hill equation is:

Y = P_O2ⁿ / (P₅₀ⁿ + P_O2ⁿ)

Rearrangement gives the Hill plot:

log[Y/(1 − Y)] = n log P_O2 − n log P₅₀

The slope is the Hill coefficient, n_H. For myoglobin n ≈ 1.0, indicating non-cooperative binding. For adult HbA, n_H is typically 2.7–3.0 over the steep mid-portion, although the theoretical maximum is 4. Thermodynamically, n_H is not the number of binding sites but an empirical measure of interaction energy between sites: n_H >1 denotes positive cooperativity, n_H =1 independent binding, and n_H <1 negative cooperativity. The more exact Adair model treats four sequential O₂-binding constants, but the Hill equation remains the clinically useful approximation for deriving standard OHDC coordinates.

Standard OHDC coordinates

Under standard conditions—temperature 37°C, pH 7.40, PCO₂ 5.3 kPa or 40 mmHg, normal 2,3-DPG, and adult HbA—the P₅₀ is approximately 3.5 kPa or 26.6 mmHg. By definition this is the PO₂ at which Hb is 50% saturated. Using n = 2.7:

At PO₂ 40 mmHg: Y = 40^2.7/(26.6^2.7 + 40^2.7) ≈ 0.75. This gives the classical mixed venous point: P_vO₂ 5.3 kPa or 40 mmHg, S_vO₂ 75%.

At PO₂ 100 mmHg: Y ≈ 0.97–0.98. This gives the classical arterial point: P_aO₂ 13.3 kPa or 100 mmHg, S_aO₂ 97–98%.

Point on OHDC	PO2	Saturation	Physiological significance
P50	3.5 kPa / 26.6 mmHg	50%	Index of Hb-O2 affinity
Mixed venous	5.3 kPa / 40 mmHg	~75%	Normal resting systemic extraction ~25%
Arterial	13.3 kPa / 100 mmHg	97–98%	Plateau region protects loading despite moderate hypoxaemia

Oxygen content and delivery

Total arterial oxygen content is the sum of Hb-bound and dissolved O₂:

C_aO₂ = (Hb × 1.34 × S_aO₂) + (0.0031 × P_aO₂)

where Hb is in g/dL, S_aO₂ is fractional, P_aO₂ in mmHg, and C_aO₂ in mL O₂/dL. The Hüfner constant is theoretically 1.39 mL/g, but 1.34 mL/g is used clinically because of dys-haemoglobins and incomplete availability. For Hb 15 g/dL, S_aO₂ 0.98 and P_aO₂ 100 mmHg: C_aO₂ ≈ (15 × 1.34 × 0.98) + 0.31 = ~20 mL/dL. Dissolved oxygen is therefore normally negligible but becomes relevant during hyperbaric oxygenation.

Systemic oxygen delivery is:

D_O2 = Cardiac output × C_aO₂ × 10

The factor 10 converts dL to L. With cardiac output 5 L/min and C_aO₂ 20 mL/dL, D_O2 ≈ 1000 mL/min. Normal oxygen consumption is approximately 250 mL/min, giving an extraction ratio of about 25%, matching the arterial-to-mixed venous saturation difference on the OHDC.

Clinical assessment and investigations

Measurement and calculation of P50

P50 is the oxygen tension at which haemoglobin is 50% saturated and is the standard quantitative descriptor of haemoglobin-O₂ affinity. In normal adult blood, standard P50 is approximately 26.6–27.0 mmHg or 3.5–3.6 kPa at pH 7.40, PCO₂ 40 mmHg, 37°C and normal 2,3-DPG. A low P50 indicates increased affinity and a left-shifted curve; a high P50 indicates reduced affinity and a right shift.

Direct P50 measurement is performed by tonometry: a blood sample is equilibrated with gas mixtures to achieve 50% oxyhaemoglobin saturation, and the corresponding PO₂ is measured. This is accurate but slow and rarely performed acutely. In practice, blood gas analysers derive P50 from measured PO₂, pH, PCO₂, temperature and either measured or calculated saturation, then correct it to standard conditions. The correction is essential: the in vivo P50 may be transiently altered by acidosis, hypercapnia, fever and 2,3-DPG, whereas standard P50 is intended to reflect intrinsic haemoglobin affinity.

The classical Severinghaus description of the oxyhaemoglobin dissociation curve provides a practical empirical relation between PO₂ and saturation under standard conditions. A commonly quoted form is:

S_O2 = (PO₂³ + 150PO₂) / (PO₂³ + 150PO₂ + 23400)

where S_O2 is fractional saturation and PO₂ is in mmHg. At PO₂ approximately 26.8 mmHg, this equation gives S_O2 close to 0.50. For individual samples, P50 may also be estimated using a Hill-type transformation: P50 = PO₂ × [(1 − S)/S]^1/n, with the Hill coefficient near the middle of the curve typically around 2.6–2.8. This is least reliable at very high saturations, because small saturation errors produce large P50 errors.

Measurement	Principle	Exam-relevant limitation
Calculated sO2 from ABG	Uses PO2, pH and assumed normal dissociation curve	Invalid with dyshemoglobins, abnormal haemoglobins, major P50 shifts
Co-oximeter saturation	Multi-wavelength spectrophotometry separates HbO2, HHb, COHb, MetHb	Requires adequate sample and calibration; lipaemia, dyes and very high bilirubin may interfere
Standard P50	P50 corrected to pH 7.40, PCO2 40 mmHg, 37°C	Reflects intrinsic affinity more than current tissue unloading conditions

Co-oximetry principles

Modern arterial blood gas analysers often incorporate co-oximetry. Unlike a two-wavelength pulse oximeter, a co-oximeter uses multiple wavelengths, commonly four or more and in some devices many dozens, to solve simultaneous absorption equations for haemoglobin species. It reports fractional oxyhaemoglobin saturation as HbO₂ divided by total haemoglobin species, and may also report functional saturation as HbO₂ divided by HbO₂ plus deoxyhaemoglobin. This distinction is central in carbon monoxide poisoning and methaemoglobinaemia.

Normal values are: carboxyhaemoglobin <2% in non-smokers, often 5–10% in smokers; methaemoglobin usually <1–2%. A PaO₂ electrode measures dissolved oxygen tension and may be normal despite critically reduced oxygen content if much haemoglobin is bound as COHb or MetHb.

Pulse oximetry: physiological and optical basis

Pulse oximetry applies the Beer–Lambert law: absorbance is proportional to the extinction coefficient, chromophore concentration and optical path length. Oxyhaemoglobin and deoxyhaemoglobin have different absorption spectra: deoxyhaemoglobin absorbs more red light at approximately 660 nm, whereas oxyhaemoglobin absorbs more infrared light at approximately 940 nm. At the isosbestic region around 805 nm, absorption is similar.

The device separates pulsatile arterial absorption from non-pulsatile venous, tissue and bone absorption by analysing the AC/DC ratio at both wavelengths. The ratio of ratios is then mapped to arterial saturation using empirical calibration in healthy volunteers, typically validated over 70–100% saturation. Accuracy under usual clinical conditions is approximately ±2–3%, but deteriorates below 70%, during low perfusion, motion artefact, severe vasoconstriction, venous pulsation, nail pigment, ambient light contamination and after intravenous dyes such as methylene blue or indocyanine green. Current anaesthetic monitoring standards, including ASA and Association of Anaesthetists guidance, require continuous pulse oximetry during anaesthesia and recovery, but the monitor must be interpreted physiologically rather than as a direct measure of oxygen content.

Important failures: dyshemoglobins and affinity shifts

Condition	Blood gas/co-oximetry finding	Pulse oximetry behaviour	Clinical implication
Carbon monoxide poisoning	PaO2 normal; COHb raised; functional oxygen content reduced; curve left-shifted for remaining sites	Falsely normal or high because COHb is interpreted partly as HbO2	Check co-oximetry. COHb half-life: room air about 4–6 h, 100% O2 about 60–90 min, hyperbaric O2 about 20–30 min. Consider hyperbaric oxygen for neurological features, myocardial ischaemia, severe acidosis, pregnancy, or COHb commonly >25% and >15% in pregnancy.
Methaemoglobinaemia	MetHb raised; chocolate-brown blood; PaO2 may be normal	Converges toward about 85% because MetHb absorbs red and infrared light similarly	Treat significant symptoms or MetHb usually >20–30% with methylene blue 1–2 mg/kg IV over 5 min; may repeat once after 30–60 min. Avoid or use caution in G6PD deficiency and serotonergic toxicity risk.
Left-shift states: alkalosis, hypothermia, low 2,3-DPG, HbF, CO	Reduced P50	SpO2 may appear reassuring	Oxygen unloading is impaired despite acceptable saturation.
Right-shift states: acidosis, hypercapnia, fever, raised 2,3-DPG	Increased P50	Lower saturation for a given PaO2	Unloading is facilitated, but arterial loading may be compromised at low alveolar PO2.

Management, pharmacology and procedures

Clinical manipulation of the oxygen-haemoglobin dissociation curve is usually indirect: altering pH, PaCO₂, temperature, red-cell 2,3-bisphosphoglycerate, dys-haemoglobin concentration, or inspired oxygen tension. For FRCA purposes, distinguish oxygen content from affinity: a left shift may improve loading but impair unloading; a right shift improves unloading but may reduce pulmonary saturation when PaO₂ is low.

2,3-BPG, stored blood and the Rappoport-Luebering shunt

In erythrocytes, 2,3-BPG is generated from 1,3-BPG by bisphosphoglycerate mutase and hydrolysed to 3-phosphoglycerate by 2,3-BPG phosphatase, bypassing ATP generation in the Rappoport-Luebering shunt. 2,3-BPG binds preferentially to deoxyhaemoglobin between the beta chains, stabilising the tense state and increasing P₅₀. Normal adult P₅₀ is approximately 26-27 mmHg at pH 7.40, 37°C and PaCO₂ 40 mmHg.

During red-cell storage, glycolysis slows, pH falls, ATP and 2,3-BPG decline. 2,3-BPG is substantially depleted by 7-14 days depending on additive solution and storage temperature, producing a left shift. After transfusion, 2,3-BPG typically regenerates within 24-72 h, faster in well-perfused recipients. Clinically, this matters most in massive transfusion, neonatal exchange transfusion, profound anaemia, or limited cardiac reserve. Contemporary trials such as ABLE, RECESS and TRANSFUSE did not show outcome benefit from routinely using fresher blood in general critical care/cardiac surgery populations, so guidelines favour standard-issue compatible red cells except for specific indications, e.g. neonatal exchange or intrauterine transfusion where fresh irradiated blood is often used.

Management of acute shift states

Problem	Mechanism	Exam-relevant management
Severe hypothermia	Left shift from reduced temperature; decreased tissue unloading despite increased Hb affinity. Oxygen analysers report at 37°C unless temperature-corrected; alpha-stat versus pH-stat is relevant during bypass.	Stage clinically: mild 32-35°C, moderate 28-32°C, severe less than 28°C. Rewarm actively when unstable: forced-air warming, warmed fluids, humidified gases; extracorporeal life support for arrest or severe instability. Avoid overinterpreting high SaO2; tissue delivery may be poor.
Therapeutic alkalosis	Left shift from raised pH and reduced PaCO2; useful only rarely to alter pulmonary vascular tone or intracranial pressure, but impairs unloading.	Avoid prolonged hyperventilation. In traumatic brain injury, Brain Trauma Foundation guidance reserves PaCO2 30-35 mmHg for short-term control of acute intracranial hypertension; prophylactic severe hypocapnia, particularly PaCO2 less than 25 mmHg, is harmful. Correct metabolic alkalosis with chloride/potassium repletion, acetazolamide 250-500 mg enteral/IV, or renal replacement in selected cases.
Methaemoglobinaemia	Fe2+ oxidised to Fe3+; reduces functional Hb and left-shifts remaining ferrous Hb. Pulse oximetry tends towards 85%; PaO2 may be normal; co-oximetry is diagnostic.	Treat if symptomatic or MetHb usually greater than 20-30%, or lower thresholds with anaemia, pregnancy, cardiac disease. Give methylene blue 1-2 mg/kg IV over 5 min; may repeat once after 30-60 min. It is reduced by NADPH-metHb reductase to leucomethylene blue. Avoid or use extreme caution in G6PD deficiency; risk haemolysis and poor response. Alternatives: ascorbic acid, exchange transfusion, hyperbaric oxygen in refractory severe cases.
Carbon monoxide poisoning	CO affinity for Hb approximately 200-250 times oxygen; forms carboxyhaemoglobin, reduces oxygen content and left-shifts residual Hb. Standard pulse oximetry is falsely normal.	Give 100% oxygen immediately. COHb half-life is about 4-6 h in air, 60-90 min on 100% O2, and 20-30 min at 2.5-3 ATA hyperbaric oxygen. Consider hyperbaric oxygen for loss of consciousness, neurological deficit, myocardial ischaemia, severe acidosis, pregnancy, or COHb greater than 25% in adults; in pregnancy many use greater than 15-20% or any fetal compromise. UHMS indications are commonly cited; evidence is mixed, but delayed neurocognitive sequelae justify aggressive treatment in high-risk cases.

Anaesthetic drugs and oxygen carriage

Volatile anaesthetics do not materially alter haemoglobin’s intrinsic oxygen affinity at clinical concentrations, but they reduce oxygen delivery by dose-dependent myocardial depression, vasodilatation and impairment of hypoxic pulmonary vasoconstriction. HPV inhibition is greatest at higher MAC and contributes to shunt during one-lung anaesthesia; intravenous agents such as propofol, etomidate and opioids generally preserve HPV better, although hypotension and hypoventilation may dominate clinically. Nitrous oxide dilutes alveolar oxygen if high concentrations are used and can worsen diffusion hypoxia unless oxygen is administered during emergence.

Several anaesthesia-related exposures create dys-haemoglobins. Desflurane, enflurane and isoflurane can react with desiccated carbon dioxide absorbents to generate carbon monoxide, especially with strong bases and very dry absorbent; prevention is by fresh absorbent, avoiding prolonged high fresh gas flows through idle machines, and using modern low-alkali absorbents. Benzocaine, prilocaine, topical local anaesthetics, nitrates, dapsone and nitric oxide can cause methaemoglobinaemia; inhaled nitric oxide requires MetHb monitoring, commonly aiming to keep MetHb 5%. Prilocaine risk rises with doses above approximately 6 mg/kg; benzocaine sprays are unpredictable and should be minimised. The practical peri-operative response to unexplained low saturation with normal PaO₂ is arterial blood gas with co-oximetry, calculation of CaO₂, and treatment of the specific dys-haemoglobin rather than blind escalation of ventilatory pressures.

Exam controversies and advanced synthesis

Bohr effect versus Haldane effect: the classical viva trap

The common Primary FRCA error is to describe carbon dioxide carriage when asked about oxygen affinity, or vice versa. The Bohr effect is the effect of CO₂ and H⁺ on the oxygen–haemoglobin dissociation curve: increased PCO₂ or reduced pH stabilises the deoxy T-state of haemoglobin, increases P50, and shifts the curve to the right, facilitating tissue unloading. At 37°C, pH 7.40 and PCO₂ 5.3 kPa, adult HbA has a P50 of approximately 3.5 kPa or 26–27 mmHg. A fall in pH from 7.40 to 7.20 may increase P50 by roughly 4–6 mmHg, although the exact value depends on temperature, 2,3-DPG and Hb species.

The Haldane effect is the reciprocal phenomenon: oxygenation status of haemoglobin determines CO₂ carriage. Deoxygenated Hb is a better buffer and forms carbamino compounds more readily; oxygenation in the lung reduces CO₂ carriage and promotes CO₂ excretion. Thus, Bohr explains why metabolically active tissues unload oxygen; Haldane explains why venous blood carries more CO₂ at a given PCO₂ and why oxygenation in the pulmonary capillary assists CO₂ elimination.

Effect	Independent variable	Dependent curve	Direction in tissues	Exam phrase
Bohr	CO2/H+	O2–Hb curve	Right shift; oxygen unloading	“CO2 affects O2 affinity”
Haldane	O2 saturation	CO2 dissociation curve	Deoxy-Hb carries more CO2	“O2 affects CO2 carriage”

The double-Bohr effect in uteroplacental exchange

Placental gas exchange is an elegant integrated test of both effects. Fetal CO₂ diffuses down its gradient into maternal blood. In the maternal intervillous space, the rise in CO₂ and H⁺ shifts the maternal HbA curve to the right, promoting O₂ unloading. Simultaneously, fetal loss of CO₂ shifts the fetal curve to the left, promoting O₂ uptake. This paired maternal right shift and fetal left shift is the double-Bohr effect.

It is reinforced structurally by fetal haemoglobin. HbF has α₂γ₂ chains, binds 2,3-DPG less avidly than HbA, and has a P50 around 19–20 mmHg compared with 26–27 mmHg for HbA. The examiner may ask whether the Haldane effect also operates in the placenta: it does. Oxygenation of fetal Hb reduces fetal CO₂ carriage, assisting fetal CO₂ unloading, while deoxygenation of maternal Hb increases maternal CO₂ uptake.

The “physiological anaemia” paradox of pregnancy

Pregnancy appears paradoxical because haemoglobin concentration falls while oxygen requirement rises. Plasma volume increases by approximately 40–50%, red cell mass by 20–30%, and Hb commonly falls by 10–20 g/L, with a nadir at 28–34 weeks. WHO defines anaemia in pregnancy as Hb 110 g/L; UK practice commonly uses <110 g/L in the first trimester and <105 g/L in the second and third trimesters. This dilution is not maladaptive: lower viscosity improves uteroplacental flow, and maternal cardiac output rises by 30–50%. Uterine blood flow increases from approximately 50–100 mL/min pre-pregnancy to 500–700 mL/min at term.

Oxygen delivery is determined by DO₂ = cardiac output × arterial oxygen content, not by Hb alone. Arterial oxygen content is approximately CaO₂ = 1.34 × Hb × SaO₂ + 0.003 × PaO₂ when PaO₂ is in mmHg. Therefore, moderate haemodilution may be compensated by increased flow, but severe anaemia is not benign. Many obstetric transfusion guidelines advise individualised red cell transfusion when Hb is 70 g/L, especially with symptoms, bleeding, cardiac disease or perioperative risk, while prioritising iron replacement when stable.

Why a left shift is not always beneficial

A left-shifted curve improves pulmonary loading at low alveolar PO₂, but oxygen transport is a two-step process: loading at the lung and unloading at the tissue. The steep portion of the curve exists to permit large changes in saturation for small changes in tissue PO₂. If affinity is too high, venous PO₂ falls and tissue extraction becomes energetically and diffusionally constrained. Normal mixed venous saturation is about 65–75%; systemic oxygen extraction is 20–30%; critical DO₂ is often quoted around 250–330 mL/min in adults, below which VO₂ becomes supply-dependent.

Clinically important left shifts include hypothermia, alkalosis, reduced 2,3-DPG, carbon monoxide poisoning and high-affinity haemoglobin variants. Stored red cells lose 2,3-DPG substantially over 1–2 weeks of storage, with regeneration after transfusion over approximately 24–72 hours. Carbon monoxide is doubly harmful: it reduces available Hb binding sites and left-shifts the remaining curve, impairing unloading despite a deceptively preserved PaO₂. Low-P50 haemoglobin variants may present with erythrocytosis because tissues sense hypoxia despite high measured SaO₂. The sophisticated answer is therefore contextual: a left shift may aid uptake in the lung or placenta, but in systemic tissue hypoxia it can worsen the final common endpoint, mitochondrial oxygen delivery.