Blood Gases, Oximetry, and Special Testing
Arterial sampling and its substitutes
Arterial blood gas analysis: the radial artery is preferred; perform a modified Allen test when indicated, use a heparinized syringe, expel air bubbles immediately (air raises PaO2 and lowers PaCO2), cap, and analyze promptly - within 15 to 30 minutes, on ice if delay is unavoidable. Results are reported at 37 °C; temperature correction is applied when the patient's temperature is abnormal.
Key derived values: PAO2 = FiO2 × (barometric pressure − 47) − PaCO2 ÷ 0.8. The alveolar-arterial gradient rises with age, V/Q mismatch, and shunt. Bicarbonate and base excess describe the metabolic component; confirm saturation by co-oximetry rather than trusting a calculated value.
Pulse oximetry is convenient, not definitive. Accuracy is roughly ±2% to ±3% between 70% and 100% saturation. Falsely high readings occur with carboxyhemoglobin; methemoglobin drives readings toward the mid-80s regardless of true saturation; poor perfusion, motion, ambient light, artificial nails, and nail polish degrade signal quality. SpO2 says nothing about ventilation - a normal saturation does not exclude hypercapnia.
Special testing to know: methacholine challenge (a 20% fall in FEV1 at a concentration of 16 mg/mL or less indicates hyperresponsiveness), six-minute walk test with continuous oximetry, maximal inspiratory and expiratory pressures, high-altitude simulation testing, and shunt studies.
Trap: titrating oxygen by oximetry alone is common but limited - confirm with arterial sampling when the clinical picture and the saturation disagree.
Worked example: FiO2 0.21, PaCO2 40 mmHg, PaO2 90 mmHg at sea level. PAO2 = 0.21 × 713 − 50 = 99.7 mmHg; A-a gradient ≈ 10 mmHg, normal for a young adult.