Oximetry, capnography and blood gas analysis
The Beer-Lambert law and how pulse oximetry and co-oximetry work and fail, infrared capnography and the capnogram, oxygen analysers, and the electrodes in a blood gas machine (pH, carbon dioxide and oxygen) with the effects of temperature and sampling errors.
By the end of this page you should be able to
- State the Beer-Lambert law and explain how a pulse oximeter derives saturation, and list its limitations.
- Distinguish pulse oximetry from co-oximetry and explain how carboxyhaemoglobin and methaemoglobin affect each.
- Explain how infrared capnography works and interpret normal and abnormal capnograms and the arterial to end-tidal gap.
- Describe the principles of the paramagnetic and fuel cell oxygen analysers.
- Describe the pH, Severinghaus and Clark electrodes, and the main errors in blood gas sampling and temperature correction.
Light absorption and pulse oximetry
The Beer-Lambert law combines two ideas: Beer's law (absorption of light is proportional to the concentration of the absorbing substance) and Lambert's law (absorption is proportional to the path length through it). Knowing how much light of a given wavelength passes through a sample lets you calculate the concentration of an absorbing substance.
Oxyhaemoglobin and deoxyhaemoglobin absorb light differently. At 660 nm (red), deoxyhaemoglobin absorbs much more; at 940 nm (infrared), oxyhaemoglobin absorbs slightly more. The isosbestic points (about 590 and 805 nm) are where both absorb equally. A pulse oximeter shines red and infrared light through a finger or earlobe, switching rapidly between them with an off phase to measure ambient light. Most light is absorbed by tissue, venous blood and non-pulsatile arterial blood; only the pulsatile (AC) component is due to arterial blood. The oximeter calculates the ratio of AC to DC absorbance at each wavelength, then the ratio of ratios (R) between red and infrared. R is converted to saturation using an empirical calibration curve derived from healthy volunteers breathing hypoxic mixtures, which is why accuracy falls below about 70 to 80%.
Limitations: it measures functional saturation among the two haemoglobins it assumes are present, not oxygen content or tension; it responds to desaturation with a lag (worse at the finger than the ear, and with poor perfusion); it needs a pulsatile signal, so it fails with vasoconstriction, hypothermia, low output and non-pulsatile flow; it is disturbed by motion, ambient light, nail varnish and some dyes (methylene blue and indocyanine green read low); it over-reads in darker skin pigmentation at lower saturations, increasing the risk of missed hypoxaemia; and venous pulsation (tricuspid regurgitation) reads low. Because of the shape of the dissociation curve, a saturation in the high 90s says little about whether arterial oxygen tension is 100 or 500 mmHg, and a patient on high inspired oxygen can be hypoventilating badly while saturation stays normal.
Dyshaemoglobins and co-oximetry
Carboxyhaemoglobin absorbs red light much like oxyhaemoglobin and hardly absorbs infrared, so a pulse oximeter counts it largely as oxyhaemoglobin: SpO2 over-reads, and a normal reading does not exclude carbon monoxide poisoning. Methaemoglobin absorbs both wavelengths roughly equally, driving R towards 1, so SpO2 drifts towards about 85% whatever the true saturation (falsely low when true saturation is high, falsely high when it is low).
A co-oximeter in a blood gas analyser lyses the sample and measures absorbance at many wavelengths, so it can calculate the fractions of oxyhaemoglobin, deoxyhaemoglobin, carboxyhaemoglobin and methaemoglobin. It reports fractional saturation (oxyhaemoglobin as a fraction of all haemoglobin). Multi-wavelength pulse oximeters can estimate carboxyhaemoglobin and methaemoglobin non-invasively but less accurately.
Capnography
Carbon dioxide absorbs infrared light strongly at about 4.3 micrometres. A capnometer passes infrared light through the gas and measures how much reaches a detector; absorption rises with carbon dioxide concentration (Beer-Lambert again), compared with a reference chamber. Mainstream analysers sit in the breathing circuit, respond fast and need no gas sampling, but are heavier and heated to prevent condensation. Sidestream analysers aspirate gas (typically 50 to 200 mL/min) to a remote cell; they are lighter and can measure other gases, but add a delay, can be blocked by water, and remove gas from the circuit, which matters in small children. Nitrous oxide absorbs near the same wavelength, and pressure broadening of the absorption band causes error that the analyser must correct. Photoacoustic and Raman methods also exist.
The time capnogram has four phases: phase I (inspiratory baseline and dead space gas), phase II (rapid upstroke as alveolar gas mixes with dead space), phase III (the alveolar plateau, slightly upsloping because units with low V/Q and long time constants empty last), and phase 0 (the inspiratory downstroke). The end-tidal value is the peak of phase III. In healthy lungs the arterial to end-tidal gap is a few mmHg; it widens with alveolar dead space (low cardiac output, pulmonary embolism, hypovolaemia, high airway pressure, emphysema).
Uses: confirming tracheal intubation (a sustained waveform over several breaths, though it can persist briefly after oesophageal intubation if the stomach contains exhaled gas); detecting disconnection, apnoea, rebreathing, bronchospasm (a shark fin with a steep phase III), malignant hyperthermia (rising carbon dioxide production), and return of spontaneous circulation in cardiac arrest (a sudden rise). A falling end-tidal value with an unchanged ventilator setting means less carbon dioxide is reaching the lungs: think cardiac output first.
Oxygen analysers
- Paramagnetic analyser: oxygen has two unpaired electrons and is attracted into a magnetic field; most other gases are weakly repelled. In the classic design, nitrogen-filled glass spheres on a dumbbell in a non-uniform field rotate as oxygen is drawn into the field; the force needed to hold them still is proportional to oxygen partial pressure. Modern versions alternate the field and detect the resulting pressure differential between sample and reference gas. Fast enough for breath-by-breath measurement; affected by water vapour, so samples are dried.
- Galvanic fuel cell: oxygen diffuses through a membrane to a gold or silver cathode where it is reduced, with a lead anode being oxidised, generating a current proportional to oxygen partial pressure without an external power source. Slow response (seconds), so used to monitor inspired oxygen; the anode is consumed, so the cell has a limited life, and it needs temperature compensation.
- Polarographic (Clark) electrode: as below; used in blood gas machines and some inspired oxygen monitors.
Blood gas electrodes
- pH electrode: a pH-sensitive glass membrane separates the sample from a buffer of known pH inside the electrode, which contains a silver/silver chloride electrode; a calomel or silver/silver chloride reference electrode completes the circuit through a salt bridge. The potential difference across the glass depends on the hydrogen ion gradient (about 61 mV per pH unit at 37 °C, from the Nernst equation). Kept at 37 °C and calibrated with two buffers.
- Carbon dioxide (Severinghaus) electrode: a modified pH electrode. Carbon dioxide diffuses through a Teflon or silicone membrane into a thin layer of bicarbonate solution, forms carbonic acid, and lowers its pH; the pH change is proportional to the logarithm of carbon dioxide tension. Slow (minutes), needs a stable temperature, and is calibrated with known gases.
- Oxygen (Clark) electrode: a platinum cathode and silver/silver chloride anode in potassium chloride electrolyte behind an oxygen-permeable membrane, with about 0.6 V applied. Oxygen is reduced at the cathode, consuming electrons, so current is proportional to oxygen tension (a polarographic cell). Needs a polarising voltage, consumes oxygen (so the sample must be stirred) and drifts as the membrane ages.
- Derived values: bicarbonate is calculated from pH and carbon dioxide using the Henderson-Hasselbalch equation; base excess is calculated from the buffer line; saturation may be measured by co-oximetry or calculated.
- Errors and temperature: air bubbles pull oxygen tension towards about 150 mmHg and carbon dioxide towards zero; delay lets leucocytes keep consuming oxygen and producing carbon dioxide (worse with leucocytosis, slowed by cooling); excess liquid heparin dilutes the sample and lowers carbon dioxide and bicarbonate. The machine measures at 37 °C. In a hypothermic patient, the true in vivo gas tensions are lower and pH higher; alpha-stat management interprets the uncorrected values (preserving the ionisation of imidazole groups and enzyme function, and cerebral autoregulation), while pH-stat corrects to the patient's temperature and adds carbon dioxide to keep a corrected pH of 7.4 (raising cerebral blood flow).
Structuring a measurement answer
- Start with the physical principle (Beer-Lambert, paramagnetism, electrochemistry), then the device, then calibration, then errors.
- For oximetry, explain the two wavelengths, the pulsatile signal and the ratio of ratios before listing limitations.
- Draw a normal capnogram with labelled phases before discussing abnormal shapes.
- For each electrode, name the anode, cathode, membrane and electrolyte, and what is measured: voltage or current.
Common traps
- Saying a pulse oximeter measures oxygen content or partial pressure.
- Saying carbon monoxide poisoning lowers SpO2; pulse oximetry over-reads.
- Confusing the electrodes: the Clark electrode measures current (oxygen), the Severinghaus electrode measures a pH-derived voltage (carbon dioxide).
- Assuming end-tidal carbon dioxide equals arterial; the gap widens with dead space.
- Calling bicarbonate on a blood gas a measured value.
Where this meets anaesthesia
- A sustained capnography trace over several breaths confirms tracheal intubation; colour change or chest movement does not.
- Pulse oximetry lags during rapid desaturation, especially with a finger probe in a cold patient.
Viva questions
Answer each one aloud, as you would to an examiner, before you open the model points.
- 1.
How does a pulse oximeter work?
Show answer
- Beer-Lambert law: absorbance proportional to concentration and path length.
- Red (660 nm) absorbed more by deoxyhaemoglobin; infrared (940 nm) slightly more by oxyhaemoglobin.
- Pulsatile component isolated; ratio of red to infrared AC/DC ratios calculated.
- Converted to saturation using an empirical calibration from volunteers; less accurate below about 70 to 80%.
- 2.
What are the limitations of pulse oximetry?
Show answer
- Needs pulsatile flow: fails with vasoconstriction, hypothermia, low output and movement.
- Measures saturation not content or tension; insensitive to hypoventilation on supplemental oxygen.
- Errors from dyshaemoglobins, dyes, nail varnish, ambient light, venous pulsation.
- Lag in response; over-reads in darker skin at lower saturations.
- 3.
What happens to the pulse oximeter reading with carbon monoxide and methaemoglobin?
Show answer
- Carboxyhaemoglobin absorbs red like oxyhaemoglobin, little infrared: SpO2 over-reads.
- Methaemoglobin absorbs both wavelengths similarly, R tends to 1, SpO2 tends to about 85%.
- Co-oximetry uses many wavelengths to measure each species and gives fractional saturation.
- Both reduce oxygen content and shift the curve left.
- 4.
How does a capnograph work?
Show answer
- Carbon dioxide absorbs infrared at about 4.3 micrometres; absorbance compared with a reference.
- Mainstream: fast, in circuit, heated, heavier. Sidestream: aspirated sample, delay, water traps, gas removed.
- Corrections for nitrous oxide overlap and pressure broadening.
- Alternatives: photoacoustic and Raman spectroscopy.
- 5.
Draw and describe a normal capnogram, then some abnormal ones.
Show answer
- Phase I baseline, II upstroke, III alveolar plateau with slight upslope, 0 inspiratory downstroke; end-tidal at the end of phase III.
- Shark fin with steep phase III in bronchospasm or obstruction.
- Raised baseline with rebreathing; sudden loss with disconnection or extubation.
- Falling end-tidal with low cardiac output or embolism; rising with hypermetabolism or after return of circulation.
- 6.
Why is end-tidal carbon dioxide lower than arterial, and when does the gap widen?
Show answer
- End-tidal gas includes alveoli that are ventilated but poorly perfused, diluting the carbon dioxide.
- Normally a gap of a few mmHg.
- Widens with alveolar dead space: low output, pulmonary embolism, hypovolaemia, high airway pressure, emphysema, lateral position.
- So end-tidal can falsely reassure about ventilation in shock.
- 7.
Describe the Clark oxygen electrode.
Show answer
- Platinum cathode, silver/silver chloride anode, potassium chloride electrolyte, oxygen-permeable membrane.
- About 0.6 V polarising voltage applied; oxygen reduced at the cathode.
- Current proportional to oxygen tension.
- Consumes oxygen so sample is stirred; temperature controlled; membrane ageing causes drift.
- 8.
How is carbon dioxide tension measured in a blood gas machine?
Show answer
- Severinghaus electrode: a pH electrode behind a carbon dioxide-permeable membrane with a bicarbonate solution.
- Carbon dioxide forms carbonic acid, lowering pH in proportion to the log of its tension.
- Slow response, needs constant temperature and gas calibration.
- Bicarbonate is then calculated from pH and carbon dioxide.
- 9.
What errors arise in arterial blood gas sampling?
Show answer
- Air bubbles: oxygen towards about 150 mmHg, carbon dioxide towards zero.
- Delay: ongoing metabolism lowers oxygen and raises carbon dioxide; worse with leucocytosis or thrombocytosis.
- Excess liquid heparin dilutes the sample, lowering carbon dioxide and bicarbonate.
- Venous contamination; temperature differences between patient and analyser.
- 10.
Explain alpha-stat and pH-stat management in hypothermia.
Show answer
- Gas solubility rises as temperature falls, so in vivo tensions are lower and pH higher than the machine at 37 °C reports.
- Alpha-stat: interpret uncorrected values; preserves imidazole ionisation, enzyme function and cerebral autoregulation.
- pH-stat: correct to body temperature and add carbon dioxide to keep corrected pH 7.4; increases cerebral blood flow.
- Choice depends on the setting, for example adult bypass against deep hypothermic arrest in children.
Practise this week
- Investigations Lab: gases, ECGs and lung function
- Primary MCQs: measurement and equipment
- Primary MCQs: quality and safety
- Primary SAQs, marked part by part
- Primary vivas
Week 3 in the other subjects
- Physiology: Carbon dioxide transport and control of breathing
- Pharmacology: Metabolism, interactions and safe prescribing
- Procedural anatomy: Neck, pharynx and larynx
Not affiliated with the Australian and New Zealand College of Anaesthetists (ANZCA). Notes are a revision aid, not a substitute for the curriculum or for clinical guidelines.