Pulse Oximetry Looks Normal After a House Fire
A 34-year-old man is brought to the emergency department after firefighters pulled him from a ground-floor bedroom fire. He is confused and combative, with soot around the nares and singed nasal hair. Vital signs are RR 30, HR 122, BP 138/84, temperature 37.4 C. EMS placed a 2 L/min nasal cannula en route, and the pulse oximeter reads 98%. An arterial sample drawn on that oxygen shows pH 7.30, PaCO2 32 mmHg, PaO2 94 mmHg, and HCO3 16 mEq/L. Co-oximetry reports carboxyhemoglobin 31%, hemoglobin 15.4 g/dL, and lactate 7.1 mmol/L. The admitting resident states that the saturation is fine and wants to place the patient on a medical floor for observation. He has no cardiac or pulmonary history and takes no medications.
1. Why is the pulse oximeter reading of 98% misleading, and roughly how much oxygen-carrying capacity has this patient lost? Show the calculation.
Read the worked answer
A conventional two-wavelength pulse oximeter cannot separate oxyhemoglobin from carboxyhemoglobin; both absorb light similarly at 660 and 940 nm, so the displayed value is a ratio of total light-absorbing hemoglobin species, not true oxyhemoglobin. The device is falsely reassuring.
Calculate arterial oxygen content: CaO2 = (1.34 x Hb x SaO2) + (0.003 x PaO2).
Because 31% of hemoglobin is bound to carbon monoxide, effective hemoglobin available for oxygen is 15.4 x 0.69 = 10.63 g/dL.
1.34 x 10.63 x 0.98 = 13.96 mL/dL, plus dissolved oxygen (0.003 x 94) = 0.28 mL/dL. Total CaO2 is about 14.2 mL/dL.
For a normal 15.4 g/dL hemoglobin fully saturated, CaO2 would be roughly 20.6 mL/dL. This patient has lost about 30% of oxygen-carrying capacity even though the saturation display reads 98%. The elevated lactate of 7.1 mmol/L is consistent with tissue dysoxia.
This is exactly the data gap standard blood gas analyzers create: most calculate SaO2 from PaO2, so they also report a normal value. Only co-oximetry reveals the true picture.
2. Based on the data, what oxygen therapy and disposition decisions do you recommend?
Read the worked answer
Treat with the highest feasible inspired oxygen immediately, not 2 L/min by cannula. Apply a tight-fitting nonrebreather at 15 L/min, or high-flow oxygen, because the half-life of carboxyhemoglobin falls from roughly 4 to 6 hours on room air to about 60 to 90 minutes on 100% oxygen.
Disposition should be critical care monitoring, not a floor bed. Indications for hyperbaric oxygen evaluation include carboxyhemoglobin greater than 25%, loss of consciousness or persistent confusion, metabolic acidosis, cardiac involvement, and pregnancy. This patient meets at least three: COHb 31%, altered mental status, and pH 7.30 with lactate 7.1.
Do not titrate oxygen down based on pulse oximetry in this situation. Use co-oximetry. Also consider concomitant cyanide toxicity from smoke inhalation: soot, altered mentation, severe lactic acidosis, and a normal PaO2 with a wide gradient; if suspected, hydroxocobalamin is the treatment. Reassess airway frequently, since inhalation injury and declining mental status may require intubation.
3. Four hours later, carboxyhemoglobin is 4% and SpO2 is 97% on 2 L/min, but the patient is still confused and combative. How do you interpret this, and what do you recommend?
Read the worked answer
The falling carboxyhemoglobin is expected and does not mean the patient is safe. The level has a short half-life on oxygen and correlates poorly with clinical severity or neurologic outcome. The persistent confusion now needs its own explanation: delayed neurologic sequelae of carbon monoxide, hypoxic brain injury, cyanide toxicity, concurrent head trauma from the fire, alcohol or drug intoxication, or developing cerebral edema.
Recommend continued high-concentration oxygen until the patient is clinically stable and acid-base status is corrected, serial neurologic examinations, a CT scan of the head, and a toxicology screen. Obtain a lactate and repeat co-oximetry. Consider hyperbaric oxygen consultation again even with a low level, on the strength of persistent neurologic deficit; the evidence is weaker for late treatment, but the decision should not be made on the number alone.
Document clearly that a normal pulse oximeter and a low carboxyhemoglobin do not equal adequate oxygen delivery in a patient with ongoing end-organ signs.