Pulse oximetry versus arterial blood gas specimens in long-term oxygen therapy.

Portable pulse oximeters are now widely available for the assessment of arterial oxygenation, and the U.S. Medicare program considers saturation readings to be acceptable substitutes for arterial PO2 in selecting patients for long-term oxygen therapy (LTOT). Current oximeters are reasonably accurate (plus or minus 4 or 5 percent of the co-oximetry value), but the clinician should be aware of several potential problems. Readings may be inaccurate in the presence of hemodynamic instability, carboxyhemoglobinemia, jaundice, or dark skin pigmentation, and also during exercise. Indicated saturation may substantially overestimate arterial PO2 if the patient is alkalemic. Pulse oximetry cannot detect hypercapnia or acidosis. For these and other reasons, pulse oximetry should not be used in initial selection of patients for LTOT, as a substitute for arterial blood gas analysis in the evaluation of patients with undiagnosed respiratory disease, during formal cardiopulmonary exercise testing, or in the presence of an acute exacerbation. Pulse oximetry is an important addition to the clinician's armamentarium, however, for titrating the oxygen dose in stable patients, in assessing patients for desaturation during exercise, for sleep studies, and for in-home monitoring.