The Saturations Don't Match

No - the number and the blood gas describe two different patients.

• A PaO2 of 92 mmHg sits on the flat upper part of the oxyhemoglobin dissociation curve, where SaO2 is roughly 96–97%. That matches the co-oximeter's directly measured 97%.
• A functional pulse oximeter calibrated against co-oximetry should track that within a few points in good perfusion. A 9-point gap (88% vs 97%) is outside normal accuracy and is not explained by the patient's acidosis (a right shift lowers saturation only slightly at this PaO2).
• Co-oximetry also rules out the classic false-low causes: carboxyhemoglobin 1.2% and methemoglobin 0.8% are both normal.

So the monitor is very likely artifact, and the arterial sample is the better reference. Two cautions before acting on that conclusion: the sample must be a genuinely arterial, appropriately handled specimen (a venous sample would show a much lower PaO2), and saturation is not oxygen delivery - at hemoglobin 7.1 g/dL this patient's oxygen content is low even at 97% saturation.

The correct first move is to validate the signal, not to chase the number with FiO2.

Causes ranked for this patient:

1. Poor peripheral perfusion - high-dose norepinephrine plus cold resuscitation has caused vasoconstriction; the finger is a marginal site.
2. Hypothermia (35.4 °C) - reduces local blood flow and slows the pulsatile signal.
3. Signal quality/probe problems - low-amplitude pleth, probe too tight or too loose, degraded LED/photodiode, ambient light, motion.
4. Venous pulsation / venous congestion - can depress the ratio-of-ratios calculation.
5. Wrong site or wrong probe type for the patient's size and perfusion.

Sequence (verify before you escalate therapy):

• Confirm the displayed pulse rate matches the ECG rate and that the pleth is genuinely pulsatile, not a wandering baseline.
• Reposition the probe; try a different finger on a warmer hand, then a forehead or ear sensor, which are less affected by peripheral vasoconstriction.
• Warm the extremity; check for a tight blood pressure cuff or restraint on that arm.
• Substitute a known-good probe to separate sensor failure from monitor failure.
• Correlate with a repeat co-oximetry sample if the discrepancy persists.

Document each substitution - the value of this work is proving which element you ruled out.

Recommendation: do not go to FiO2 0.80 on the basis of an unverified 88%.

• The measured SaO2 is 97% on FiO2 0.40. Escalating oxygen for an artifact exposes the patient to absorption atelectasis, possible oxygen toxicity with prolonged high FiO2, and - more importantly - misdirects attention from the real problem, which is likely perfusion and oxygen-carrying capacity, not oxygenation.
• The safe framing: keep the current FiO2 while the sensor is validated, and let the team make ventilator changes off the co-oximeter/ABG rather than the pulse oximeter.
• If a verified, well-perfused reading really were 88%, that would be a different conversation and FiO2 escalation would be appropriate - the decision hinges on the verification, not on the number alone.

What to tell them about the technology:

• Pulse oximetry estimates functional saturation - the fraction of available hemoglobin binding sites occupied by oxygen. It says nothing about hemoglobin concentration, cardiac output, or tissue delivery. This patient is anemic at 7.1 g/dL; a reassuring saturation can coexist with poor oxygen delivery.
• It is unreliable in low-perfusion states and can be fooled by dyshemoglobins (excluded here) and by motion or poor probe contact.
• Ask that the discrepancy and the verification steps be documented in the record so the next shift does not re-litigate the same number.