A patient with a fingertip pulse oximeter reading 92% and a patient with an arterial blood gas reporting PaO₂ 58 mmHg are being described with two different measurements of the same thing — blood oxygenation — and each captures something the other misses. In Indian respiratory practice, where a ₹600 oximeter is ubiquitous and an arterial blood gas is not, confusion between these numbers leads to real clinical errors: patients started on long-term oxygen therapy on the basis of a single SpO₂ reading taken with cold fingers, patients denied LTOT because their spot-check SpO₂ read 89% and the clinician did not recognise that their true arterial saturation was well below that. This article unpacks what SpO₂, PaO₂, and SaO₂ each measure, where they agree, where they diverge, and when each one is the right number to act on.
The audience: respiratory therapists, physicians in training, pharmacy and home-care staff prescribing oxygen, and engaged patients who want to understand their own readings. The bias throughout is toward the Indian clinical context — the equipment commonly available, the skin tones of the patient population, and the realities of where an arterial puncture is actually feasible.
The three measurements
The three numbers describe oxygen in three physically different places.
PaO₂ is the partial pressure of oxygen dissolved in the plasma of arterial blood, measured in mmHg (or kPa in some units). It is obtained from an arterial blood gas (ABG) sample, typically a radial artery puncture. PaO₂ reflects the driving pressure that pushes oxygen from plasma onto haemoglobin and from the capillary into tissue. Normal adult PaO₂ at sea level is 80–100 mmHg. PaO₂ falls with age; a reasonable predictor is PaO₂ ≈ 100 − (age/3) mmHg for adults breathing room air.
SaO₂ is the arterial oxygen saturation of haemoglobin, expressed as a percentage of haemoglobin molecules carrying oxygen. It is measured by CO-oximetry on an arterial blood sample — the same sample that yields PaO₂, analysed with a multi-wavelength photometer that can distinguish oxyhaemoglobin, deoxyhaemoglobin, carboxyhaemoglobin (COHb), and methaemoglobin (MetHb). Normal adult SaO₂ is 95–100% at sea level.
SpO₂ is peripheral capillary oxygen saturation, measured non-invasively by a pulse oximeter that shines red (660 nm) and infrared (940 nm) light through a tissue bed (finger, toe, earlobe) and computes the ratio of pulsatile absorption at each wavelength. It approximates SaO₂ but is not identical to it. Normal adult SpO₂ is 95–100% at sea level.
The arithmetic relating them:
- PaO₂ and SaO₂ are linked by the oxyhaemoglobin dissociation curve (see below).
- SaO₂ and SpO₂ should agree within ±2% when peripheral perfusion is good, no confounding haemoglobin species is present, and the oximeter is a clinical-grade device with proper calibration.
The oxyhaemoglobin dissociation curve
At the bedside, the single most important physiological concept is the sigmoid relationship between PaO₂ and SaO₂. The curve is not linear. A PaO₂ of 100 mmHg gives SaO₂ ~98%; a PaO₂ of 60 mmHg gives SaO₂ ~90%; a PaO₂ of 40 mmHg gives SaO₂ ~75%. The inflection near PaO₂ 60 mmHg is the clinically critical zone — this is where small PaO₂ changes produce large SaO₂ changes, and where most LTOT prescription thresholds sit.
The curve shifts with physiological state:
- Right shift (reduces haemoglobin’s affinity for O₂ at any given PaO₂): acidosis, hypercapnia, hyperthermia, elevated 2,3-DPG. A right-shifted curve means that at a given PaO₂, SaO₂ is lower — harder to saturate but easier to offload oxygen at tissue.
- Left shift (increases affinity): alkalosis, hypocapnia, hypothermia, carboxyhaemoglobinaemia, fetal haemoglobin, stored blood. A left-shifted curve at a given PaO₂ reads a higher SaO₂, but tissue offloading is impaired.
The clinical lesson: a patient with septic shock and lactic acidosis reading SpO₂ 92% has a true PaO₂ that is lower than the curve predicts for a well-compensated patient. Trending SpO₂ alone in a metabolically unstable patient is unreliable.
What a pulse oximeter catches and misses
A pulse oximeter infers saturation from the differential absorbance of red and infrared light by oxyhaemoglobin vs deoxyhaemoglobin across a pulsatile tissue bed. The design choice of using only two wavelengths introduces specific blind spots.
Carboxyhaemoglobin (COHb) absorbs red light almost identically to oxyhaemoglobin. A pulse oximeter reads COHb as O₂Hb. In a smoker with COHb of 10%, an SpO₂ of 98% may correspond to a true SaO₂ of 88%. In carbon monoxide poisoning, SpO₂ is catastrophically misleading — this is a scenario where ABG with CO-oximetry is mandatory, not optional. In the Indian context, CO exposure from indoor biomass burning and from vehicle exhaust in dense urban settings is under-recognised as a confounder.
Methaemoglobin (MetHb) absorbs red and infrared light in proportions that make the oximeter converge toward ~85% regardless of true SaO₂. Patients exposed to dapsone, topical anaesthetics, nitrates, aniline dyes (occasionally in Indian textile exposure), or with G6PD deficiency and oxidant stress can have clinically significant MetHb.
Poor peripheral perfusion. The oximeter needs a pulsatile arterial signal. Cold hands, vasoconstriction (shock, sympathetic tone), peripheral vascular disease, or Raynaud’s all reduce the signal. The device may still display a number, but with low perfusion index (PI) the reading has much wider error bars — some machines display PI, which is worth checking. In a patient with PI below 0.4, many oximeters produce readings that deviate from true SaO₂ by 5 percentage points or more.
Motion artefact. Tremor, shivering, and positioning movement corrupt the pulsatile waveform. Clinical-grade oximeters use signal-extraction technology (Masimo SET, Nellcor OxiMax) to discriminate signal from motion; many consumer fingertip oximeters do not. A reading taken during a patient’s first 30 seconds of placement, while the sensor is still settling, is unreliable.
Nail polish, henna, and artificial nails. Dark nail polish, fresh henna, and opaque artificial nails alter light transmission. In Indian practice, henna is common; it produces a reading bias that varies with henna density and wavelength and is best avoided by placing the sensor sideways across the finger, on the earlobe, or removing polish before a diagnostic reading.
Pulse oximeter accuracy and skin pigmentation
A substantial body of literature from 2020 onward documented a systematic bias in pulse oximeter readings for patients with darker skin pigmentation — the devices over-read SpO₂ (read higher than true SaO₂) more frequently in Black patients than in white patients, and the discrepancy is clinically material. The US Food and Drug Administration issued a safety communication in 2022 acknowledging this bias and convened an advisory committee in 2022 and 2024 to address it. The mechanism is still being characterised; melanin absorbance at the relevant wavelengths is implicated, as are device calibration datasets that historically under-represented dark-skinned subjects.
Indian skin tones span the Fitzpatrick IV–VI range across most of the population. The direct applicability of North American dark-skin studies to Indian populations is not perfectly established — but the mechanism implies that Indian patients, particularly those with darker skin, are also susceptible to over-reading. The clinical implication: an Indian patient reading SpO₂ 92% may truly sit at SaO₂ 88–90%. For LTOT qualification decisions made at 88% SpO₂, this matters directly.
Practical adjustments in Indian practice: when the decision is borderline, use an ABG rather than relying on a single SpO₂ reading; use multiple readings across different fingers and conditions; consider the earlobe as a sampling site (it is less affected by peripheral vasoconstriction); and be explicit with the patient about the margin of uncertainty.
Normal values at sea level vs altitude
Normal reference ranges for all three numbers are altitude-dependent. At sea level, a healthy adult has SpO₂ 96–99%, SaO₂ 95–100%, and PaO₂ 80–100 mmHg. At 2,000 m altitude (Darjeeling, Shimla, Ooty), expected SpO₂ falls to 92–95% in acclimatised healthy adults. At 3,500 m (Leh), the expected range is 86–92%. At 5,000 m, 75–85%. A PaO₂ of 55 mmHg that triggers LTOT consideration at sea level is physiologically closer to normal for a resident of Leh — but Leh residents with PaO₂ 55 mmHg still have elevated mortality risk and many are appropriately on LTOT. Altitude modifies the baseline, not the threshold at which supplementation helps. The Indian Chest Society has not published an altitude-specific modifier; physicians at altitude generally apply the sea-level thresholds with clinical judgement.
When ABG is clinically mandatory, not optional
Pulse oximetry is adequate for most monitoring. It is inadequate as the sole basis for some decisions.
ABG is mandatory when:
- Initiating long-term oxygen therapy. GOLD 2024 and the Indian Chest Society both stipulate that LTOT qualification is based on PaO₂ measured by ABG, not on SpO₂ (GOLD Report). Using SpO₂ <88% as a surrogate for PaO₂ <55 mmHg is acceptable in initial triage and in settings where ABG is unavailable, but a definitive LTOT prescription — especially one intended to be reimbursed or continued for years — should be anchored to a measured PaO₂.
- Evaluating acid-base status. Whenever CO₂ retention, metabolic acidosis, or mixed disturbances are clinically suspected, SpO₂ provides no information at all — only PaCO₂ and pH from an ABG do.
- Carbon monoxide or methaemoglobin exposure suspected. SpO₂ is unreliable in these scenarios.
- Critically ill patient with discordance between clinical appearance and SpO₂. A patient who looks cyanotic with a reported SpO₂ of 96% is not having a reassuring clinical picture — they are having a probe or calibration problem.
- Pre-titration of home NIV or supplemental O₂ in patients with hypercapnic respiratory failure. CO₂ retention cannot be monitored without ABG (or transcutaneous CO₂, which is clinical-grade and not widely available in Indian home-care).
SpO₂ alone is sufficient for:
- Routine monitoring of a known hypoxaemic patient on stable LTOT.
- Ambulatory desaturation studies (the 6-minute walk test uses SpO₂).
- Titrating oxygen flow in the acute ward where ABGs are obtained intermittently.
- Screening well patients or triaging in primary care.
Practical Indian-context reality
The equipment reality in India: fingertip pulse oximeters are sold without regulatory traceability for ₹500–₹1,500. Many are not ISO 80601-2-61 compliant. Clinical-grade handheld units (Masimo Rad, Nellcor PM10N, Nonin) sit in the ₹15,000–₹60,000 range and live in tertiary-care ICUs, pulmonology OPDs, and some home-care service providers. A primary-care clinic prescribing LTOT typically has access to one of the consumer-grade devices and no ABG facility.
The pragmatic workflow most competent Indian pulmonologists follow: screen with SpO₂, confirm borderline readings with a good-quality oximeter and multiple samples across perfusion conditions, refer for ABG when initiating LTOT or when any clinical concern suggests hypercapnia. In tier-2 and tier-3 cities, ABG availability is concentrated at district hospitals and larger private hospitals. In rural settings, ABG is often genuinely unavailable, and LTOT is prescribed on SpO₂ alone with explicit acknowledgement that the prescription is provisional until ABG can be obtained.
CGHS reimbursement for LTOT typically requires documentation of qualifying arterial hypoxaemia. For CGHS beneficiaries this generally means an ABG report; for ECHS (military retirees) the practice varies by hospital. PMJAY does not at present provide a national scheme for home LTOT equipment, leaving most rural patients paying out-of-pocket for oxygen concentrators and therefore without a reimbursement lever that would force rigorous ABG documentation.
The result is a care gap: patients with genuinely qualifying hypoxaemia may go undiagnosed because their SpO₂ was measured once with cold fingers and read 91%; patients with provisional SpO₂-based prescriptions may not have their therapy properly titrated because no follow-up ABG ever happens. Both errors are addressable — the first by recognising the 2–5 percentage-point uncertainty of spot SpO₂ and the second by insisting on an ABG at 30–90 days of LTOT initiation.
Closing: which number when
The shorthand a respiratory therapist carries to the bedside:
- PaO₂ tells you the driving pressure — use it for LTOT qualification, acid-base interpretation, and any ICU-level decision.
- SaO₂ is what the haemoglobin is actually carrying — use it when MetHb or COHb is suspected and you need a CO-oximetry-derived saturation.
- SpO₂ is a continuous, cheap, non-invasive surrogate for SaO₂ — use it for monitoring, titration, and screening, and treat every number below 93% as a question to investigate rather than a number to dismiss.
Patients and carers benefit from understanding that their home pulse oximeter is a screening and monitoring tool, not a diagnostic final answer. If the number matters for a decision — starting oxygen, changing prescription, evaluating a symptomatic patient — the number needs context, repetition, and sometimes an ABG. Consult your pulmonologist when the numbers and symptoms disagree, or when a single reading would change a major clinical decision.
Primary references that inform clinical practice in this area: GOLD Report 2024; Indian Chest Society guidelines on LTOT 2017; ATS/ERS 2020 statement on home oxygen therapy; FDA Safety Communication on pulse oximeter accuracy 2022; Sjoding et al. NEJM 2020 on racial bias in pulse oximetry.