Two Oximeters, One Patient: Which Saturation Do You Trust?
A 71-year-old woman with a 45-pack-year smoking history is admitted with an acute COPD exacerbation. She is alert and using accessory muscles. Vital signs: HR 112/min, BP 92/58 mmHg, RR 26/min, temperature 38.1°C. Her hands are cool and mottled, and she wears dark polish on both index fingers. A right index finger probe reads SpO2 86% with a weak, irregular plethysmographic waveform; an earlobe probe placed one minute later reads 91% with a strong, steady waveform. An arterial blood gas drawn on 2 L/min nasal cannula 10 minutes ago shows pH 7.36, PaCO2 61 mmHg, PaO2 58 mmHg, HCO3 34 mEq/L. The nurse asks you which oximeter number to chart and act on.
1. Which SpO2 value should be used for clinical decisions, and what in this scenario makes the other value unreliable?
Read the worked answer
Use the earlobe reading of 91%.
Reasoning steps:
- Check signal quality first. The finger tracing has a weak, irregular plethysmograph - a low-perfusion signal. Without a good waveform, the numeric value is not interpretable at all. The earlobe waveform is strong and steady.
- Look for local factors at the finger: cool, mottled skin (vasoconstriction from shock-ish BP 92/58 and fever), plus dark nail polish, which attenuates the red/infrared signal. Any one of these can cause a falsely low reading; together they make the 86% untrustworthy.
- Cross-check against the ABG. PaO2 58 mmHg corresponds to roughly 87–89% on the oxyhemoglobin dissociation curve. The earlobe value of 91% is within the few-percent error expected of pulse oximetry; 86% is further off and in the wrong direction for the perfusion problem described.
- Recognize the limitation: agreement with the ABG is supportive, not proof. Pulse oximetry still cannot detect hypercarbia, and a normal-ish SpO2 never rules out ventilatory failure in this patient.
2. Using the blood gas, is this patient hypoxemic, and what saturation range should you titrate to? Does the PaCO2 of 61 mmHg demand an immediate ventilator change?
Read the worked answer
Hypoxemia: yes. PaO2 58 mmHg is below the 60 mmHg threshold for hypoxemia; she is not adequately oxygenated on 2 L/min.
Target: 88–92% SpO2 (PaO2 roughly 55–60 mmHg). She has chronic hypercapnia - HCO3 34 with pH 7.36 means the CO2 is chronically compensated, not acute. In a known CO2 retainer, uncontrolled high-flow oxygen risks worsening hypercarbia and absorption atelectasis, so the goal is the lowest oxygen that keeps her in the 88–92% band, not a "normal" 98%.
The CO2 itself does not require an immediate ventilatory change. pH 7.36 with a chronic bicarbonate indicates her baseline. The actionable problem is that she is below her target saturation: increase delivery (e.g., titrate nasal cannula upward with reassessment, or a 28% Venturi mask for tighter control) and recheck.
Escalation triggers: falling pH, rising CO2 above her baseline, worsening mental status, or inability to hold 88–92% with low-flow oxygen - that is when NIV is considered, not on the CO2 number alone.
3. What must be documented about these oximetry measurements, and when should you recheck?
Read the worked answer
Document enough that the next clinician can judge reliability:
- Site and probe type (right index finger vs. earlobe) and that the finger signal was inadequate.
- Signal quality, not just the number: weak/irregular plethysmograph on the finger, strong on the earlobe.
- Interfering factors: cool mottled extremities, poor perfusion, dark nail polish.
- The concurrent ABG values and the oxygen device/flow at the time, so the SpO2 and PaO2 are interpreted as a pair.
- Which value was used and why, so the trend line stays internally consistent.
Recheck timing: after any change in oxygen delivery, and at minimum with vital signs (usually every 1–2 hours in an acute exacerbation), or sooner if her work of breathing, mental status, or respiratory rate changes. Because the plethysmograph was poor, every subsequent reading should be accompanied by a waveform check - a number without a waveform is not a measurement.