A Saturation That Does Not Match the Arterial Oxygen Tension
A 62-year-old man is 18 hours out from an elective total knee arthroplasty performed under spinal anesthesia with a supraglottic airway. Since surgery he has used an over-the-counter benzocaine throat spray every 1-2 hours for a sore throat. He is alert, oriented, speaking in full sentences, and reports only incisional pain. Vital signs: RR 18, HR 98, BP 142/86 mmHg, temperature 37.2 C. Pulse oximetry reads 86-88% on two different fingers while on 2 L/min nasal cannula; the plethysmographic waveform is low-amplitude but regular. An arterial sample drawn by the resident appears unusually dark. ABG on 2 L/min: pH 7.42, PaCO2 39 mmHg, PaO2 88 mmHg, HCO3 25 mEq/L. Co-oximetry has been ordered but results are pending.
1. The SpO2 reads 87% while the PaO2 is 88 mmHg. Explain what that discordance means, build a differential, and state what you do in the next 15 minutes.
Read the worked answer
What the discordance means
At a PaO2 of 88 mmHg, the oxyhemoglobin dissociation curve predicts an SpO2 near 96-97%. The measured 87% is a saturation gap of roughly 9-10 points, and because it reproduced on two fingers it is not explained by a single bad probe.
Differential
- Dyshemoglobinemia. Benzocaine is a classic cause of acquired methemoglobinemia, and this man has had repeated doses. Methemoglobin absorbs light at both wavelengths the standard two-wavelength oximeter uses, so the device drifts toward a fixed reading near 85% that does not move with FiO2. Carboxyhemoglobin and sulfhemoglobin behave similarly.
- Artifact. Poor perfusion, low pulse amplitude, motion, ambient light, dark nail polish, IV dyes.
- Misleading blood gas. Most analyzers report a calculated SaO2 derived from PaO2 and pH, not a measured one; a normal calculated SaO2 is fully compatible with a low true saturation.
Next 15 minutes
- Keep oxygen applied and get co-oximetry with SpMet and COHb; do not let a rising FiO2 be the only intervention.
- Move and reposition the probe, confirm perfusion, and note the pleth quality.
- Notify the physician of the saturation gap and the benzocaine exposure with time of last dose.
- Label SpO2 in the chart as unreliable and titrate against measured co-oximetry and clinical status.
- Stop the throat spray and keep the remaining bottle so the exposure time is documented.
2. Co-oximetry returns an SpMet of 24%. The patient is alert, comfortable, and his lactate is 1.8 mmol/L. What do you recommend, and how does your answer change if he becomes confused with a lactate of 5 mmol/L?
Read the worked answer
Asymptomatic, SpMet 24%, normal lactate
This is observation territory, not antidote territory. Remove the offending agent, continue oxygen, obtain serial co-oximetry, and monitor mental status and lactate. Untreated methemoglobin reductase activity is slow, so the level may take 12-24 hours to fall; a rising trend or new symptoms changes the plan.
If he becomes confused with a lactate of 5 mmol/L
That is end-organ hypoxia and an indication for methylene blue 1-2 mg/kg IV over about 5 minutes. Before giving it, check G6PD status: in G6PD deficiency methylene blue is ineffective and can cause hemolysis, and the alternatives are ascorbic acid, exchange transfusion, or hyperbaric oxygen. Watch for two traps: methylene blue itself is an oxidant at high doses and can cause a falsely low SpO2 reading, and benzocaine is lipophilic, so levels can rebound as it redistributes from tissue.
Monitoring Trend SpMet by co-oximetry, not pulse oximetry; document neuro status, lactate, and hemodynamics; anticipate rebound for several hours after the last exposure.
3. The bedside nurse asks why you are not simply increasing his oxygen. What is your explanation, and what exactly do you document at handoff?
Read the worked answer
Explanation for the nurse
His hemoglobin is carrying oxygen poorly because a fraction of it is oxidized to methemoglobin, which cannot bind oxygen and shifts the dissociation curve. Adding FiO2 raises dissolved oxygen and PaO2 but cannot convert methemoglobin back; the monitor will keep reading about 85% no matter what we do, so a higher flow is not a treatment and a reassuring number is not a reassurance. The real targets are SpMet by co-oximetry, mental status, and lactate.
Documentation
- Probe type and site, pleth quality, and whether the reading reproduced elsewhere.
- Simultaneous SpO2 and ABG values, plus the calculated saturation gap.
- Co-oximetry results with timestamps and the analyzer used.
- Exposure history: agent, dose frequency, last dose.
- Interventions, including removal of the agent and any antidote.
- An explicit statement that SpO2 is unreliable for titration and what is being used instead.
Handoff Use a structured format: this patient has methemoglobinemia; the number on the monitor is not his saturation; use measured co-oximetry; the offending agent has been removed; here are the criteria for methylene blue; here is the rebound window.