Guide — Labs & Diagnostics
Co-oximetry & Dyshemoglobinemias
A standard pulse oximeter sees only two hemoglobin species and assumes nothing abnormal is present — so it can read reassuringly normal while oxygen delivery is collapsing. This guide explains how laboratory co-oximetry measures the true hemoglobin species, exposes a saturation gap, and guides treatment of carbon monoxide poisoning, methemoglobinemia, and sulfhemoglobinemia.
9 min read · Labs & Diagnostics
Written by Apex Respiratory Editorial Team
Educational use only. This material supports respiratory therapy education and exam review. It is not medical advice and is not a substitute for clinical judgment, institutional protocols, or physician orders. Always follow facility policies and current provider orders, and verify calculations independently before clinical use.
Overview
A pulse oximeter is a fast, noninvasive workhorse, but it is also a limited instrument: it reports oxygen saturation on the assumption that the only hemoglobin species present are oxyhemoglobin and deoxyhemoglobin. When an abnormal hemoglobin — a dyshemoglobin — is in the blood, that assumption breaks, and the SpO₂ can read reassuringly normal while tissue oxygen delivery is severely impaired. The classic example is carbon monoxide poisoning, where both the SpO₂ and the PaO₂ look fine.
The instrument that does see the abnormal species is the laboratory co-oximeter, run on a blood gas sample. Understanding why these two devices disagree — and recognizing the bedside clues that should prompt a co-oximetry blood gas — is a core respiratory care competency and a recurring board topic.
Two wavelengths vs. four
The difference between a pulse oximeter and a co-oximeter comes down to how many wavelengths of light each one uses, and therefore how many hemoglobin species each one can tell apart.
- Pulse oximeter — two wavelengths. A standard pulse oximeter uses just two wavelengths of light, 660 nm red and 940 nm infrared. With only two data points it can distinguish oxyhemoglobin from deoxyhemoglobin and nothing more. It reports functional saturation and assumes no abnormal hemoglobins are present.
- Co-oximeter — four or more wavelengths. A laboratory co-oximeter uses four or more wavelengths and directly measures the fractions of oxyhemoglobin, deoxyhemoglobin, carboxyhemoglobin (COHb), and methemoglobin (MetHb). It reports fractionalsaturation — the true SaO₂ — because it accounts for the abnormal species the pulse ox is blind to.
That distinction is the whole point: when a dyshemoglobinemia is suspected, the correct test is a blood gas with co-oximetry, not the pulse oximeter on the finger.
The saturation gap
The single most useful screening concept is the saturation gap. When the pulse-ox SpO₂ is meaningfully higher than the co-oximeter SaO₂ — a gap greater than about 5% — suspect a dyshemoglobinemia. The pulse ox is reading high precisely because it cannot see the abnormal species and is counting bound, non-functional hemoglobin as if it were carrying oxygen. A gap that size should prompt you to look for carbon monoxide exposure or an oxidizing-drug history and to draw a co-oximetry sample rather than trusting the finger probe.
The three dyshemoglobins
Three abnormal hemoglobin species account for nearly all clinically relevant dyshemoglobinemias. Each fools the pulse oximeter in its own way, and each has its own treatment path.
| Species | Typical cause | Pulse-ox effect | Treatment |
|---|---|---|---|
| Carboxyhemoglobin (COHb) | Carbon monoxide — smoke inhalation, exhaust, faulty heaters | Falsely normal SpO₂; PaO₂ also normal | 100% oxygen; hyperbaric oxygen for severe cases |
| Methemoglobin (MetHb) | Oxidizing drugs — benzocaine and local anesthetics, dapsone, nitrates/nitrites | SpO₂ trends toward a fixed ~85% plateau | Methylene blue (avoid in G6PD deficiency — use ascorbic acid or exchange transfusion) |
| Sulfhemoglobin | Some sulfonamides — rarer, irreversible oxidation | Cyanosis; may be misreported as methemoglobin | No antidote; resolves only as red cells turn over |
Carboxyhemoglobin (CO poisoning).Carbon monoxide binds hemoglobin with about 240 times the affinity of oxygen, left-shifting the dissociation curve and crippling oxygen delivery. Because COHb absorbs light much like oxyhemoglobin at the pulse-ox wavelengths, the SpO₂ reads falsely normal — and the PaO₂ is also normal, since CO does not change dissolved oxygen. Diagnosis requires co-oximetry: COHb above roughly 3% in a nonsmoker, up to about 10–15% in smokers, with severe poisoning above about 25%. Treat with 100% oxygen, which shortens the CO half-life from hours to under an hour, and use hyperbaric oxygen for severe cases. The “cherry-red” skin sign is late and unreliable — do not wait for it.
Methemoglobin.Oxidizing drugs — topical benzocaine and other local anesthetics, dapsone, and nitrates and nitrites — convert hemoglobin iron to the ferric state, which cannot carry oxygen and also left-shifts the curve. Because MetHb absorbs both wavelengths nearly equally, the pulse ox trends toward a fixed reading of about 85% regardless of the true saturation. The clues are cyanosis that does not respond to oxygen and chocolate-brown blood. Cyanosis appears at roughly 1.5 g/dL of MetHb, around 10–15%. Treat significant methemoglobinemia with methylene blue — but methylene blue is ineffective and can cause hemolysis in G6PD deficiency, where ascorbic acid or exchange transfusion is used instead.
Sulfhemoglobin.A rarer, irreversible oxidation — associated with some sulfonamides — produces cyanosis with no specific antidote; it resolves only as red cells turn over. Co-oximeters may misreport sulfhemoglobin as methemoglobin, so it should be considered when a suspected methemoglobinemia fails to respond to methylene blue.
What the RT does with it
The recurring theme at the bedside is a mismatch between reassuring numbers and a sick patient. The respiratory therapist's job is to recognize that mismatch and act on it.
- Draw a co-oximetry blood gas when the picture does not fit. If the SpO₂ and PaO₂ look normal but the patient is symptomatic, or if cyanosis does not improve on oxygen, the finger probe is not the right test — send a co-oximetry sample.
- Deliver 100% oxygen for suspected CO poisoning. High-FiO₂ oxygen drives down the CO half-life immediately while you arrange co-oximetry and hyperbaric oxygen evaluation. Do not let a normal SpO₂ talk you out of treating.
- Recognize methemoglobinemia from its signature.An SpO₂ stuck near the ~85% plateau, chocolate-brown blood, and cyanosis that does not lift with oxygen point to MetHb — and you should flag the G6PD caution before methylene blue is given.
Common Pitfalls
- Trusting a “normal” SpO₂ in smoke inhalation or CO exposure. Carboxyhemoglobin reads like oxyhemoglobin on the pulse ox, so a normal SpO₂ does not rule out carbon monoxide poisoning. Suspect it from the exposure history, not the saturation.
- Assuming a normal PaO₂ rules out impaired oxygen delivery. Carbon monoxide does not change dissolved oxygen, so the PaO₂ stays normal even when most hemoglobin is bound and useless. Oxygen content and delivery, not PaO₂, are what suffer.
- Giving methylene blue in G6PD deficiency.In G6PD deficiency methylene blue is both ineffective and capable of triggering hemolysis — use ascorbic acid or exchange transfusion instead.
- Mistaking the MetHb 85% plateau for true desaturation. The pulse ox parks near 85% because methemoglobin absorbs both wavelengths equally, not because the patient is truly at 85% — only co-oximetry gives the real number.
Board Exam Pearls
- A pulse oximeter uses two wavelengths (660 nm and 940 nm); a co-oximeter uses four or more and measures the true hemoglobin species.
- Carbon monoxide makes the SpO₂ falsely normal — and the PaO₂ is normal too, so co-oximetry is the test.
- Methemoglobin pulls the SpO₂ toward a fixed ~85% plateau, with chocolate-brown blood and oxygen-unresponsive cyanosis.
- A saturation gap greater than 5% between SpO₂ and SaO₂ suggests a dyshemoglobin.
- Methylene blue is contraindicated and ineffective in G6PD deficiency — use ascorbic acid or exchange transfusion.
FAQ
Why does the pulse ox read normal in carbon monoxide poisoning?
Carboxyhemoglobin absorbs light at nearly the same wavelengths as oxyhemoglobin, so a two-wavelength pulse oximeter cannot tell them apart and counts the bound, useless hemoglobin as if it were carrying oxygen. The SpO₂ therefore reads falsely normal, and the PaO₂ is normal too because carbon monoxide does not change dissolved oxygen. Both routine numbers look reassuring while oxygen delivery is severely impaired, which is why co-oximetry is required to make the diagnosis.
What is a saturation gap?
A saturation gap is the difference between the pulse-ox SpO₂ and the true SaO₂ measured by laboratory co-oximetry. When the SpO₂ sits meaningfully higher than the co-oximeter SaO₂ — a gap greater than about 5% — suspect a dyshemoglobinemia. The pulse ox reads high because it cannot see the abnormal hemoglobin species, whereas the co-oximeter measures all of them directly.
Why does methemoglobinemia make the SpO₂ sit around 85%?
Methemoglobin absorbs the 660 nm red and 940 nm infrared light nearly equally, which drives the ratio the pulse oximeter uses toward a fixed value. As methemoglobin rises, the SpO₂ trends toward a plateau of roughly 85% and stops tracking the true saturation. A reading stuck near 85% that does not improve with oxygen, alongside chocolate-brown blood and cyanosis, is the classic clue.
When is methylene blue unsafe?
Methylene blue is the treatment for significant methemoglobinemia, but it is ineffective and can trigger hemolysis in patients with G6PD deficiency. In that setting, ascorbic acid or exchange transfusion is used instead. Always flag the G6PD caution before methylene blue is given.
Related Resources
Sources
- American Association for Respiratory Care. AARC Clinical Practice Guideline: Blood Gas Analysis and Hemoximetry. Respir Care. 2013;58(10):1694-1703.
- Kacmarek RM, Stoller JK, Heuer AJ. Egan's Fundamentals of Respiratory Care. 12th ed. Elsevier; 2021. Analysis and monitoring of gas exchange; hemoximetry.