Comprehensive Review of Pulse oximetry and its role in Anaesthesia & critical care
Reading File
Finding Sources
Finding Sources
Searching PubMed
pulse oximetry anaesthesia critical care
Reading File
Reading File
Reading File
Reading File
Finding Sources
Finding Sources
Reading File
Finding Sources
Finding Sources
Searching Images
pulse oximetry principle Beer Lambert law hemoglobin absorption wavelength
Reading File
Loading Image
Loading Image
Reading File
Finding Sources
Reading File
Finding Sources
Searching PubMed
pulse oximetry accuracy racial disparity hidden hypoxemia
Reading File
Searching PubMed
pulse oximetry hidden hypoxemia skin pigmentation
Reading File
Loading Image
This comprehensive review covers all major aspects of pulse oximetry relevant to anaesthesia and critical care practice, synthesised from leading textbooks and current literature. Here's what was covered:
Structure of the review:
- Historical Background — ASA 1986 mandate, WHO/WFSA standards, closed-claims evidence
- Physiologic Fundamentals — O₂ content equation, haemoglobin species, the oxydissociation curve
- Principles of Operation — Lambert-Beer law, AC/DC signal processing, the R-ratio and calibration curve (with diagrams), transmittance vs. reflectance modes
- Anaesthesia Applications — Mandatory monitoring, MAC sedation, endobronchial intubation detection, SpO₂ targets
- Photoplethysmography — PVI for fluid responsiveness, Perfusion Index, goal-directed fluid therapy
- Limitations & Errors — Full error table, dysfunctional haemoglobins in depth (COHb, MetHb, SHb), perfusion-dependent failure, accuracy at low saturations, racial bias/skin pigmentation (emerging evidence)
- Critical Care Role — ICU monitoring, oxygenation targets, SvO₂/ScvO₂, remote monitoring (post-COVID)
- Multi-wavelength Co-oximetry — SpCO, SpMet, SpHb; pulse spectroscopy
- Cerebral NIRS — Cardiac surgery, carotid endarterectomy
- Probe Selection — Site-specific advantages; forehead preference in vasoconstriction
- Clinical Pearls — 10 high-yield practice points
Act like a professor in Anesthesia & critical care - nowTeach me pulse oximetry topic with More details
Checking Date & Time
Finding Sources
Finding Sources
Finding Sources
Reading File
Reading File
Finding Sources
Loading Image
Finding Sources
Finding Sources
Reading File
Loading Image
Loading Image
Loading Image
Loading Image
Excellent. All the key diagrams are in. Now I'll deliver the full professor-style lecture.
🎓 LECTURE: Pulse Oximetry — A Complete Professor-Level Teaching Session
Department of Anaesthesia & Critical Care
Welcome. Today we are going to master pulse oximetry — not just as a gadget you clip onto a finger, but as a physiological instrument with deep physics, well-defined failure modes, and advanced hemodynamic applications. By the end of this session you will be able to explain every number and waveform it generates, predict exactly when it will lie to you, and use it at an expert level in theatre and intensive care.
PART 1 — WHY THIS MONITOR MATTERS: A HISTORICAL PERSPECTIVE
Before pulse oximetry became routine in the mid-1980s, unrecognised hypoxaemia during anaesthesia was a major cause of preventable death and brain injury. The ASA closed-claims database repeatedly showed that respiratory events were the single largest source of adverse anaesthetic outcomes. Cyanosis — the clinical sign of hypoxia — is not detectable until SpO₂ drops below roughly 80%, meaning the patient is already critically hypoxic by the time you see it.
Takuo Aoyagi, a Japanese engineer, developed the concept of pulse oximetry in 1972, exploiting the fact that pulsatile arterial blood could be distinguished from static tissue optically. The first commercial devices appeared in the early 1980s. By 1986, the ASA mandated it for all anaesthetics. The WHO and WFSA subsequently adopted it as a global minimum monitoring standard.
Think about that: this single monitor, costing less than $200 today, transformed anaesthetic safety in a way that no drug ever has.
PART 2 — THE ESSENTIAL PHYSIOLOGY YOU MUST OWN
2.1 Oxygen Delivery — Why SaO₂ Is Everything
Oxygen delivery to tissues (DO₂) is:
DO₂ = CO × CaO₂
And CaO₂ (arterial oxygen content) is:
CaO₂ = (1.34 × Hb × SaO₂) + (0.0031 × PaO₂)
- The Hüfner constant (1.34 mL/g) is the O₂-carrying capacity of haemoglobin per gram. Theoretically 1.39, but experimentally 1.31–1.37 because of small amounts of non-functional Hb species in real blood.
- The dissolved O₂ term (0.0031 × PaO₂) is negligible under normal conditions — roughly 0.3 mL/dL at a PaO₂ of 100 mmHg. It becomes relevant only during hyperbaric oxygen therapy.
The critical teaching point: You can have a normal SpO₂ and still have catastrophically low oxygen delivery if:
- The patient is severely anaemic (low Hb — the multiplier)
- Cardiac output is critically low
- Dysfunctional Hb species are present (COHb, MetHb — which are "counted" by the oximeter but cannot carry O₂)
SpO₂ tells you the saturation fraction — it says nothing about the denominator (Hb) or the multiplier (CO).
2.2 Haemoglobin Species — Know All Five
| Species | Can Carry O₂? | Normal % | Clinical Relevance |
|---|---|---|---|
| Oxyhaemoglobin (O₂Hb) | ✅ | ~97% | The functional molecule |
| Deoxyhaemoglobin (deO₂Hb) | ✅ (unloaded) | ~2–3% | Increases in hypoxia |
| Carboxyhaemoglobin (COHb) | ❌ | 1–3% (smokers up to 10%) | CO poisoning — falsely elevates SpO₂ |
| Methaemoglobin (MetHb) | ❌ (very high affinity, won't release) | < 1% | Drug-induced — pulls SpO₂ toward 85% |
| Sulphaemoglobin (SHb) | ❌ | < 1% | Irreversible; pulls SpO₂ toward 85% |
Functional SaO₂ (what pulse oximetry measures):
= O₂Hb ÷ (O₂Hb + deO₂Hb) × 100
Fractional SaO₂ (what a lab co-oximeter measures — the gold standard):
= O₂Hb ÷ (O₂Hb + deO₂Hb + COHb + MetHb + SHb) × 100
A smoker with COHb of 10%: lab fractional SaO₂ might be 89%, but the pulse oximeter reports 99% — because it ignores the COHb. This is a life-threatening false negative.
2.3 The Oxyhaemoglobin Dissociation Curve — Master This
Fig 37.2 — Miller's Anesthesia 10e: The oxyhaemoglobin dissociation curve and its shifting factors
The curve is sigmoid (S-shaped) for good physiological reasons:
The flat upper portion (PaO₂ > 70 mmHg):
- A large fall in PaO₂ (e.g., 150 → 100 → 70 mmHg) causes only a tiny fall in SpO₂
- This protects against minor ventilation impairment
- Clinical trap: SpO₂ cannot detect hyperoxia. On 100% FiO₂, a patient might have PaO₂ of 500 mmHg — you cannot see this on the pulse oximeter. In neonates and premature infants, this is how retinopathy of prematurity develops
- The pulse oximeter will not warn you when PaO₂ rises dangerously high on supplemental oxygen
The steep portion (PaO₂ < 60 mmHg):
- Small falls in PaO₂ cause large falls in SaO₂ — the "cliff edge"
- At the inflection point around PaO₂ 60 mmHg → SpO₂ ≈ 90% — this is the critical threshold
- Below this, the patient falls off the cliff rapidly
P50 = PaO₂ at which haemoglobin is 50% saturated. Normal = 26 mmHg. This is the standard measure of curve position.
Hb-O₂ dissociation curve with physiological shifting factors — Rosen's Emergency Medicine
Right shift (↑P50 — more O₂ released to tissues):
- ↓pH (acidosis), ↑PaCO₂ (hypercarbia), ↑temperature, ↑2,3-DPG
- This is the Bohr effect — exercising muscle creates an acid environment that promotes O₂ unloading. Elegant physiology.
Left shift (↓P50 — less O₂ released to tissues):
- ↑pH (alkalosis), ↓PaCO₂, ↓temperature, ↓2,3-DPG, fetal Hb, COHb
- During hypothermic bypass: the curve shifts left — haemoglobin clings to oxygen and doesn't release it to tissues easily. Metabolic rate is also reduced, so this is less of a problem — but it matters when interpreting SvO₂
MetHb and SHb shift the curve left — they have extremely high O₂ affinity and won't release O₂ to tissues. This is why methaemoglobinaemia causes tissue hypoxia even when "SpO₂" looks acceptable.
PART 3 — THE PHYSICS: HOW THE MACHINE ACTUALLY WORKS
3.1 The Beer–Lambert Law
The physical foundation is the Beer–Lambert Law:
I_trans = I_in × e^(−ε × C × D)
Where:
- I_trans = transmitted light intensity
- I_in = incident light intensity
- ε = extinction coefficient (how strongly the solute absorbs light at that wavelength — unique for each molecule)
- C = concentration of solute
- D = path length through the solution
If you know ε (from published absorption spectra), D (path length through the tissue), and measure I_trans/I_in, you can calculate C (concentration of the solute).
For multiple solutes, you need measurements at multiple wavelengths — at least as many wavelengths as solutes you want to measure. A standard two-wavelength pulse oximeter can only distinguish two Hb species. A lab co-oximeter uses 4–7 wavelengths. Rainbow multi-wavelength oximeters use up to 12.
3.2 The Two Wavelengths — Why 660 nm and 940 nm?
Differential absorption of OxyHb vs DeoxyHb — Morgan & Mikhail 7e, Fig 6–2
Look at this absorption spectrum carefully:
- At 660 nm (red): DeO₂Hb absorbs far more than O₂Hb → deoxygenated blood absorbs red light heavily → appears blue/cyanotic to the naked eye
- At 940 nm (infrared): O₂Hb absorbs more than deO₂Hb → oxygenated blood absorbs more infrared
- There is an isosbestic point around 805 nm where both species absorb equally — at this wavelength, absorption is independent of oxygen saturation
The wavelengths 660 nm and 940 nm are chosen because the difference in extinction coefficients between O₂Hb and deO₂Hb is maximal at these points — maximising the signal-to-noise ratio of the measurement.
Now look at the full five-species absorption spectra:
Fig 37.3A/B — Miller's Anesthesia 10e: Full absorption spectra of all five Hb species. Panel B shows the two operating wavelengths (dashed lines) where COHb and MetHb interference becomes evident
Panel B is the clinically critical one. Look at the 660 nm line:
- COHb (red) and O₂Hb (blue) have very similar extinction coefficients → the oximeter cannot tell them apart → CO poisoning is invisible to a standard pulse oximeter
- MetHb (orange) has significant absorption at both 660 nm and 940 nm → it corrupts both channels equally → R approaches 1.0 → SpO₂ reads 85% regardless of true saturation
3.3 The AC/DC Trick — How It Isolates Arterial Blood
This is the most elegant part of the physics. The probe emits light through perfused tissue. What does that light pass through?
- Skin, fat, bone, muscle (constant, non-pulsatile)
- Venous blood (constant, non-pulsatile under normal conditions)
- Capillary blood (largely constant)
- Arterial blood (pulsatile with every heartbeat)
The oximeter divides the light absorption signal into:
- DC component = the constant background (all non-arterial tissue + venous blood)
- AC component = the pulsatile change caused by each arterial pulse
By taking the ratio of AC to DC at each wavelength, the DC background cancels out. What remains is a signal representing only the arterial blood.
The R ratio is then calculated:
R = (AC₆₆₀ / DC₆₆₀) ÷ (AC₉₄₀ / DC₉₄₀)
This ratio is then converted to SpO₂ using the manufacturer's internal empirical calibration curve:
Fig 37.5 — Miller's Anesthesia 10e: Calibration curve. The critical point — when R = 1 (equal absorption at both wavelengths), SpO₂ reads exactly 85%
How the calibration curve is built: Volunteers breathe progressively hypoxic gas mixtures while simultaneous SpO₂ (pulse oximeter) and SaO₂ (arterial blood gas co-oximetry) are measured. The R:SaO₂ relationship is mapped across the range 70–100% and stored in the device. The FDA requires root mean square accuracy ≤ 3.0% over this range.
Critical implication: Below 70%, the device is extrapolating outside its calibration range. Accuracy is completely unreliable below 70% SaO₂. This is also why the machine was never calibrated to detect hyperoxia — nobody made volunteers breathe 100% O₂ to map that range.
3.4 Hardware: What the Probe Contains
A standard finger probe contains:
- Two LEDs (Light Emitting Diodes) at 660 nm and 940 nm
- One photodiode (photodetector) on the opposite side of the finger
- A microprocessor cycles each LED on and off in sequence. When both are off, ambient light is measured and subtracted — this is how the oximeter rejects overhead lighting artefact
Transmittance mode (standard): LEDs and detector are on opposite sides of the tissue (finger, toe, earlobe). Light passes through the tissue.
Reflectance mode (forehead probe): LEDs and detector are on the same side. The detector captures backscattered light. Used when transmittance sites are inaccessible or vasoconstricted.
PART 4 — THE OUTPUT: WHAT THE MONITOR DISPLAYS AND WHAT IT REALLY MEANS
4.1 The SpO₂ Number
SpO₂ = peripheral oxygen saturation estimated by pulse oximetry. It estimates functional SaO₂.
Accuracy under normal conditions: ±2–3% (FDA-mandated RMSE ≤ 3.0%)
What this means in practice: An SpO₂ of 95% means the true SaO₂ is somewhere between 92% and 98% with 95% confidence. When you are treating the patient at the borderline, this uncertainty matters.
Important SpO₂ landmarks to memorise:
| SpO₂ | Approximate PaO₂ | Clinical Meaning |
|---|---|---|
| 100% | > 150 mmHg | Hyperoxia possible — cannot detect |
| 97–98% | ~100 mmHg | Normal |
| 95% | ~80 mmHg | Acceptable minimum most adults |
| 93% | ~70 mmHg | Low — action warranted |
| 90% | ~60 mmHg | Critical threshold — "cliff edge" |
| 85% | ~50 mmHg | Severe hypoxaemia — OR — MetHb/SHb artefact |
| 80% | ~45 mmHg | Cyanosis clinically visible |
| < 70% | < 40 mmHg | Device accuracy collapses |
The relationship between SpO₂ and PaO₂ is not fixed — it depends on where the patient sits on their own dissociation curve, which is shifted by pH, temperature, PaCO₂, and 2,3-DPG.
4.2 The Plethysmographic Waveform — Your Second Data Stream
The photoplethysmograph (PPG) waveform is the pulse oximeter's second output and is underused by most clinicians.
Fig 37.6 — Miller's Anesthesia 10e: Photoplethysmogram demonstrating respiratory waveform variation. At time B, ventilation stops — the variation disappears. This is the physiological basis of PVI and fluid responsiveness prediction
What the waveform shows:
- Amplitude = pulse pressure / peripheral vascular tone. Tall waves = good perfusion. Low amplitude = vasoconstriction or low stroke volume
- Regularity = rhythm. Irregular waveform → arrhythmia
- Respiratory variation (ΔPOP) = the amplitude varies with each breath during mechanical ventilation because of cyclical changes in venous return and stroke volume. This is the plethysmography variability index (PVI).
4.3 PVI and Fluid Responsiveness — Advanced Application
The physiological mechanism:
During positive pressure inspiration, intrathoracic pressure rises → venous return falls → right ventricular preload falls → after a lag of 2–3 heartbeats → left ventricular stroke volume falls → pulse pressure falls → PPG amplitude falls.
During expiration, the reverse happens.
If the patient's ventricle is preload-dependent (on the steep part of the Frank-Starling curve), this respiratory swing in stroke volume is large — high PVI → fluid responsive.
If the ventricle is preload-independent (on the plateau), the respiratory swing is small — low PVI → not fluid responsive.
PVI formula:
PVI (%) = [(PPG_max − PPG_min) ÷ PPG_max] × 100
Threshold: PVI > 13–15% generally predicts fluid responsiveness in mechanically ventilated patients
Prerequisites for valid PVI measurement:
- Controlled mechanical ventilation (not spontaneous breathing — patient effort overrides the signal)
- Sinus rhythm (arrhythmias corrupt the respiratory vs. cardiac signal separation)
- Tidal volume ≥ 8 mL/kg (smaller tidal volumes generate smaller respiratory swings — false negatives)
- No severe vasodilation (septic shock can produce venous pulsations that corrupt the DC component)
Goal-directed therapy using PVI in major abdominal surgery has been shown to improve outcomes — reduced postoperative complications, shorter hospital stay. — Miller's Anesthesia 10e, p. 5461
4.4 Perfusion Index (PI)
PI = AC amplitude ÷ DC component × 100 (%)
It is a continuous, non-invasive estimate of peripheral vascular tone.
- Normal PI: 1–10%
- Low PI (< 1%): peripheral vasoconstriction — shock, hypothermia, high-dose vasopressors, patient cold and clammy
- High PI (> 4%): peripheral vasodilation — warm sepsis, epidural, sympatholysis, high ambient temperature
Clinical uses:
- Predicting hypotension after spinal anaesthesia: a PI that rises after spinal block onset indicates vasodilation before blood pressure falls — gives a 1–2 minute warning
- Monitoring vasopressor effect: PI should rise as vasopressors take effect and redirect blood to periphery (actually reflects reversal of peripheral shutdown)
- Identifying poor probe signal: PI < 0.3% → SpO₂ reading unreliable
PART 5 — LIMITS OF THE TECHNOLOGY: WHERE IT FAILS YOU
5.1 The "85% Trap" — Two Scenarios Where You Must Think
When SpO₂ reads exactly or near 85%, think of three things:
- True SaO₂ of 85% (the patient is severely hypoxaemic)
- Methaemoglobinaemia — MetHb absorbs equally at 660 and 940 nm → R = 1 → machine reads 85%
- Optical shunt — probe is misplaced and direct LED light reaches the photodetector without passing through tissue → also reads 85%
How to distinguish: Apply 100% O₂. If SpO₂ rises toward 100%, it was true hypoxaemia. If SpO₂ stays stubbornly at 85% despite high FiO₂, suspect MetHb or probe fault. Check ABG with co-oximetry.
5.2 Carbon Monoxide Poisoning — The Invisible Killer
COHb absorbs 660 nm light almost identically to O₂Hb. At 940 nm, COHb absorbs virtually nothing.
Result: The oximeter sees COHb as O₂Hb → SpO₂ appears near normal.
Example: Patient with 40% COHb → fractional SaO₂ is ~57%, but the oximeter reads ~97%.
You must never use pulse oximetry alone to rule out CO poisoning. Always use ABG with co-oximetry (or bedside multi-wavelength oximetry with Rainbow technology) in:
- House fires
- Enclosed space events
- Unexplained altered consciousness
- Any patient with headache, nausea and flu-like symptoms
5.3 Venous Pulsations — An Underappreciated Error
The AC/DC separation works because veins are assumed non-pulsatile. This breaks down in:
- Severe tricuspid regurgitation — the right atrial backpressure transmits to peripheral veins, making them pulsatile
- Tight probe placement — occludes venous outflow, creating pulsatile venous pressure
- Trendelenburg position with forehead probe — dependent venous congestion produces venous pulsation
- Distributive shock — extreme vasodilation produces arteriovenous shunting where venous blood becomes pulsatile
Effect: The device includes venous (desaturated) blood in its "arterial" signal → falsely low SpO₂ reading
5.4 Motion Artefact
Motion creates unpredictable changes in the light path through tissue, generating AC signals unrelated to arterial pulsation. The oximeter's algorithm interprets this as a low-perfusion or low-saturation state.
Modern devices use advanced signal processing (e.g., Masimo SET — Signal Extraction Technology) that employs adaptive filtering to separate genuine arterial pulsation from motion noise. Masimo's algorithm builds a reference signal from motion using a third detector and subtracts it. This is why "Masimo" oximeters perform significantly better in patients who are moving, shivering, or being transported.
During shivering in PACU, a standard oximeter may alarm continuously and give false SpO₂ values. This is clinically important because it leads to alarm fatigue.
5.5 Low Perfusion States
At systolic BP < 80 mmHg, the AC component of the pulse oximetry signal diminishes critically — the device cannot identify a reliable pulse → absent or inaccurate readings.
Solutions:
- Move to a central site: earlobe or forehead probe — the arterial supply to these areas is less catecholamine-responsive than fingers
- Earlobe: superficial temporal artery territory — relatively vasodilation-resistant
- Forehead (reflectance probe): supratrochlear and supraorbital arteries — similarly preserved in shock
- In patients on high-dose vasopressors (noradrenaline, vasopressin), finger probes often fail — switch to earlobe or forehead
5.6 Complete Summary of Artefacts
| Source | Mechanism | Effect on SpO₂ |
|---|---|---|
| CO poisoning | COHb absorbs 660 nm like O₂Hb | Falsely ↑↑ (near normal) |
| Methaemoglobinaemia | Equal absorption both channels → R→1 | Pulled toward 85% |
| Sulphaemoglobinaemia | Similar to MetHb | Pulled toward 85% |
| Optical shunt | Direct LED-to-detector light | Reads 85% |
| Methylene blue | Peak absorption 668 nm (near HHb) | Marked transient ↓ to ~65% |
| Indocyanine green (ICG) | Mild 660 nm absorption | Mild ↓ |
| Hypotension / low flow | AC signal lost | Absent reading or ↓ |
| Motion | Corrupted AC signal | Variable, usually ↓ |
| Venous pulsations | Venous blood enters AC signal | ↓ |
| Black/dark nail polish | Absorbs 660 nm | ↓ (usually <2%) |
| Dark skin (SaO₂ < 80%) | Melanin alters optical path | Falsely ↑ (hidden hypoxaemia) |
| IABP (balloon pump) | Extra mechanical pulsations | ↑ (falsely elevated) |
| Sickle cell disease | Elevated COHb from haem turnover | May ↑ during vaso-occlusive crisis |
| Severe anaemia | Reduced AC signal | ↓ in hypoxia |
| Infrared navigation (neurosurgery) | External IR source corrupts 940 nm | ↓ or signal loss; shield with foil |
| Continuous-flow LVAD | No pulsatility → no AC signal | Device fails entirely |
| Fetal Hb | Not a significant error — accuracy preserved | Minimal effect |
5.7 The Racial Bias Issue — Modern Critical Awareness
This is a recently spotlighted clinical problem with major ethical implications.
The mechanism: At SaO₂ > 80%, skin melanin does not significantly affect SpO₂ accuracy. At SaO₂ < 80%, melanin alters the optical properties of the tissue in a way that makes the oximeter overestimate true saturation. The calibration curves were developed using predominantly light-skinned volunteers.
The consequence: Black patients with respiratory failure have a significantly higher prevalence of occult (hidden) hypoxaemia — meaning they are genuinely hypoxaemic (true SaO₂ below threshold) but their SpO₂ reads acceptably. This leads to delayed recognition, delayed treatment, and worse outcomes.
Studies during the COVID-19 pandemic quantified this: Black patients were more likely to receive delayed eligibility determination for supplemental oxygen therapy based on SpO₂ targets.
What to do: In any patient with darker skin pigmentation who has unexplained deterioration or in whom clinical findings don't match the SpO₂ reading — check an ABG with co-oximetry. Do not rely on SpO₂ alone.
PART 6 — PULSE OXIMETRY IN SPECIFIC CLINICAL SETTINGS
6.1 Intraoperative Anaesthesia
Standard of care for every anaesthetic, including MAC sedation. No contraindications.
During MAC/sedation — the critical concept:
- Pulse oximetry detects only 50% of apnoea episodes detected by capnography
- SpO₂ drops an average of 45.6 seconds after apnoea begins (because of oxygen reserve from pre-oxygenation and supplemental O₂)
- This delay is dangerous — by the time SpO₂ falls, significant hypercapnia has already occurred
- Lesson: In sedation cases, always use capnography alongside pulse oximetry. SpO₂ is a late indicator of respiratory depression when supplemental O₂ is being given
Endobronchial intubation:
- Pulse oximetry frequently fails to detect right mainstem intubation if the patient is pre-oxygenated on 100% FiO₂
- The non-ventilated left lung still has O₂ from pre-oxygenation — SpO₂ stays normal for several minutes
- Capnography is more sensitive (waveform unchanged in endobronchial intubation but auscultation and asymmetric chest rise are key)
Intraoperative oxygen targets:
- Most adult patients: SpO₂ 95–100% (avoid hyperoxia in COPD, neonates)
- Premature neonates: 91–95% (above 95% risks retinopathy of prematurity)
- Single-lung ventilation: SpO₂ ≥ 92% generally acceptable; tolerate lower if unavoidable
6.2 Post-Anaesthesia Care Unit (PACU)
The PACU is where residual anaesthesia + opioid analgesia + positioning combine to produce insidious hypoxaemia.
- Continuous SpO₂ monitoring is mandatory
- Shivering commonly causes motion artefact — use Masimo-technology device or earlobe probe
- Aim SpO₂ ≥ 95% before discharge to ward (or ≥ 92% in COPD)
- A persistently low SpO₂ in PACU — think: airway obstruction, residual neuromuscular block, opioid-induced respiratory depression, pneumothorax, PE, atelectasis
6.3 ICU
Continuous monitoring: Pulse oximetry provides real-time SpO₂ trend — alerts clinicians to acute deterioration, tube displacement, mucus plugging, pneumothorax.
Oxygenation targets in critical care:
| Condition | Target SpO₂ |
|---|---|
| General critically ill | 94–98% |
| COPD / chronic type II respiratory failure | 88–92% |
| ARDS (conservative oxygenation) | 88–95% |
| Post-cardiac arrest (avoid hyperoxia) | 94–98% — hyperoxia worsens neurological outcome |
| Premature neonate | 91–95% |
| Carbon monoxide poisoning | 100% (100% FiO₂ mandatory — SpO₂ unreliable, but maximise O₂ to displace CO) |
AVOID liberal oxygenation in the ICU — multiple RCTs (including the ICU-ROX and OXYGEN-ICU trials) have shown that hyperoxaemia is associated with increased mortality. SpO₂ > 98% on supplemental O₂ may indicate dangerous PaO₂ levels. If in doubt, check an ABG.
The SvO₂ connection:
Mixed venous oxygen saturation (SvO₂) from the pulmonary artery catheter, or ScvO₂ from the CVC, represents the balance between O₂ delivery and consumption:
SvO₂ = SaO₂ − VO₂ / (1.34 × Hb × CO)
Normal: 65–80%
- SvO₂ < 65% → extraction increased → shock (hypovolemic, cardiogenic, distributive) or ↑ VO₂ (sepsis, pain, shivering)
- SvO₂ > 80% → either excess delivery OR failure to extract (distributive shock with mitochondrial dysfunction, cytotoxic hypoxia, left-to-right shunt)
The relationship between SpO₂ (arterial) and SvO₂ (venous) frames the entire oxygen delivery-consumption physiology picture.
6.4 Prehospital and Transport
In transport, motion artefact is a constant problem. Use of Masimo or Nellcor devices with advanced signal processing is preferred. The earlobe probe is more robust in transit than finger probes.
In air ambulance (altitude), PaO₂ at cabin altitude (typically 1500–2400 metres) is reduced — SpO₂ may be 3–5% lower than at ground level. Plan supplemental O₂ accordingly.
6.5 Neonatal Screening — Critical Congenital Heart Disease (CCHD)
The US Secretary of Health and Human Services has mandated universal newborn pulse oximetry screening for critical congenital heart disease before discharge.
Protocol: SpO₂ measured in the right hand (pre-ductal) and either foot (post-ductal). Positive screen if:
- SpO₂ < 95% in either limb, OR
-
3% difference between right hand and foot (indicates right-to-left ductal shunting)
Sensitivity for CCHD: ~75% for suspected lesions, ~58% for unsuspected. When combined with antenatal ultrasound and newborn examination, up to 92% of critical CHD lesions are identified.
PART 7 — ADVANCED AND EMERGING TECHNOLOGIES
7.1 Multi-Wavelength Pulse CO-Oximetry (Masimo Rainbow)
Using up to 12 wavelengths, this technology non-invasively measures:
| Parameter | What It Measures | Limitation |
|---|---|---|
| SpO₂ | Standard arterial oxygen saturation | Standard accuracy |
| SpCO | Carboxyhaemoglobin | Not precise enough to replace lab co-oximetry; useful for screening |
| SpMet | Methaemoglobin | Accurate even during hypoxia (newer devices) |
| SpHb | Total haemoglobin | Reasonable bias ±1 g/dL; unreliable in low perfusion; limited data in Hb 6–10 g/dL range |
Clinical role of SpHb: Continuous intraoperative haemoglobin monitoring to reduce unnecessary transfusion — or to catch hidden major haemorrhage in real time. In major trauma, vascular, and liver surgery it has genuine utility, but SpHb should trigger further ABG sampling rather than be used as a standalone transfusion trigger.
7.2 Cerebral Near-Infrared Spectroscopy (NIRS / rSO₂)
This uses reflectance-mode NIRS at the forehead to measure regional cerebral oxygen saturation (rSO₂). It measures a weighted average (~75% venous, ~25% arterial) of tissue oxygenation in the frontal cerebral cortex.
Normal rSO₂: 60–75%
Key uses:
- Cardiac surgery on CPB: detect cerebral desaturation during aortic cross-clamping, hypotension, or embolism. Intervention guided by rSO₂ ≥ 50% or not < 20% of baseline
- Carotid endarterectomy: detect cerebral ischaemia during carotid cross-clamping — guides need for shunting
- Detection of malpositioning/malfunction in ECMO
- Patients with continuous-flow ventricular assist devices (where conventional SpO₂ is impossible)
7.3 Pulse Spectroscopy (Emerging)
Uses hundreds of wavelengths across the visible and near-infrared spectrum. Early results show:
- Accurate SaO₂ determination
- Reliable COHb and MetHb assessment during both normoxia and hypoxia
- Potential to fully replace ABG co-oximetry for saturation measurements in future
PART 8 — WHAT PULSE OXIMETRY CANNOT TELL YOU
Never forget these gaps — this is where clinical disaster occurs:
| What You Want to Know | Can SpO₂ Tell You? | What You Need Instead |
|---|---|---|
| Is ventilation adequate? | ❌ | Capnography, ABG PaCO₂ |
| Is there hyperoxia? | ❌ | ABG PaO₂ |
| Is oxygen delivery adequate? | ❌ | SpO₂ + Hb + cardiac output |
| Is there CO poisoning? | ❌ | ABG co-oximetry |
| Is MetHb present? | Partially (fixed at 85%) | ABG co-oximetry |
| Is tissue using oxygen? | ❌ | SvO₂/ScvO₂, lactate |
| Acid-base status | ❌ | ABG |
| Are the lungs ventilated bilaterally? | Often ❌ | Auscultation, ETCO₂, CXR |
PART 9 — THE PROFESSOR'S CLINICAL SCENARIOS
Scenario 1. A patient in the ICU post-liver transplant on noradrenaline 0.3 mcg/kg/min. SpO₂ probe on the left index finger reads 88% and is alarming. Nurse is about to increase FiO₂ further. What do you do?
The answer: The patient is on a vasopressor — fingers are vasoconstricted. This is likely a perfusion artefact. First, check the waveform — is there a good plethysmographic trace with a recognisable peak? If the trace is poor, the number is meaningless. Move the probe to the earlobe or forehead. If the SpO₂ is still low on a good waveform trace, then take an ABG. Do not blindly increase FiO₂ based on a noisy signal.
Scenario 2. A fire rescue patient is brought in with suspected smoke inhalation. SpO₂ reads 98%. He is confused and has a headache. Do you reassure him?
The answer: Absolutely not. His SpO₂ reading is almost certainly falsely elevated due to COHb. CO causes confusion and headache. Send ABG with co-oximetry immediately. Start 100% high-flow O₂ now. Do not wait for ABG results to start treatment.
Scenario 3. During laparotomy, your patient's SpO₂ drops from 99% to 85% suddenly and does not respond to increasing FiO₂. The surgeon has just injected methylene blue to test for anastomotic leak. What is happening?
The answer: Methylene blue has a peak light absorption at 668 nm — very close to the deoxy channel (660 nm). The oximeter interprets this as massive deoxygenation and SpO₂ crashes to ~65–85%. This is a well-known, transient artefact lasting 1–2 minutes until methylene blue distributes and is cleared. The patient is almost certainly not hypoxic. Check the surgical field, ventilate normally, and wait. Inform the team.
Scenario 4. A 25-year-old fitness instructor with known G6PD deficiency was given prilocaine topical cream before a procedure. Now SpO₂ reads 85% and is fixed there. 100% O₂ makes no difference. Diagnosis?
The answer: Prilocaine (and benzocaine, dapsone) causes acquired methaemoglobinaemia — the SpO₂ is being pulled toward 85% regardless of true saturation. Confirm with ABG co-oximetry (will show MetHb%). Treat with methylene blue 1–2 mg/kg IV — this activates the NADPH pathway and rapidly reduces MetHb to Hb. Note the paradox: methylene blue (which itself transiently drops SpO₂) is the treatment for methaemoglobinaemia. Use ABG, not pulse oximetry, to guide treatment.
SUMMARY FRAMEWORK — THE PROFESSOR'S ONE-PAGE MENTAL MODEL
PULSE OXIMETRY: LAYERED MENTAL MODEL
LAYER 1 — PHYSICS
Beer-Lambert Law → Lambert-Beer extinction at 660 nm & 940 nm
AC (pulsatile) / DC (static) separation → isolates arterial signal
R = (AC660/DC660) ÷ (AC940/DC940) → calibration curve → SpO₂
LAYER 2 — PHYSIOLOGY
SpO₂ estimates SaO₂ (functional, not fractional)
CaO₂ = 1.34 × Hb × SaO₂ + 0.0031 × PaO₂
SaO₂ is shaped by the oxyHb dissociation curve
Curve shifts: CADET for right shift (↑CO₂, Acidosis, ↑2,3-DPG, Exercise/Temp)
LAYER 3 — WHAT IT GIVES YOU
SpO₂ → oxygenation (not ventilation, not delivery, not tissue use)
PPG waveform → perfusion, rhythm, PVI
PI → vascular tone
Heart rate → from pulsatile signal
LAYER 4 — WHEN IT LIES
COHb → falsely HIGH (reads normal in CO poisoning)
MetHb / SHb → locked at 85%
Optical shunt → locked at 85%
Methylene blue → acutely drops to ~65%
Low perfusion → absent/low reading
Dark skin < 80% SaO₂ → falsely HIGH (hidden hypoxaemia)
Venous pulsation → falsely LOW
LAYER 5 — WHAT IT CANNOT DO
Ventilation ✗ Hyperoxia ✗ Hb concentration ✗
Tissue oxygenation ✗ CO poisoning ✗ Acid-base ✗
Sources:
- Miller's Anesthesia, 2-Volume Set, 10e — Chapter 37, pp. 5449–5473 (definitive reference)
- Morgan & Mikhail's Clinical Anaesthesiology, 7e — Chapter 6, pp. 222–225
- Barash, Cullen & Stoelting's Clinical Anaesthesia, 9e — Chapter 30, pp. 2491–2492
- Rosen's Emergency Medicine, 10e — Chapter 10 (methaemoglobinaemia)
- Gallagher JJ & Adamski JL — Accuracy of Pulse Oximetry in Non-white Patients, Crit Care Nurs Clin North Am 2025 [PMID: 40849175]
- Coutrot M et al. — Perfusion Index in Anaesthesia & Critical Care, Anaesth Crit Care Pain Med 2021 [PMID: 34687923]
We've covered everything from the Beer-Lambert law to racial equity in oximetry. Do you want to go even deeper on any specific layer — waveform interpretation, fluid responsiveness protocols, the advanced haemoglobin disorders, or how to approach a specific clinical scenario?
This is a shared conversation. Sign in to Orris to start your own chat.