An evaluation of prolonged oximetric data acquisition.

Data derived from pulse oximetry has inherent limitations, one of which is artifactual desaturation caused by patient movement. Perioperative patterns of oxygen desaturation were studied for a mean duration of 67 hours in eight young patients following corrective spinal surgery. Pulse oximetry data were relayed to a computer using Satmaster, a program which permits storage, retrieval, signal evaluation and statistical analysis of oximetry data. Desaturation episodes were mild, of short duration and their infrequent occurrence was not increased during intravenous morphine infusion. Retrospective identification of contemporaneous artifactual changes in signal amplitude permitted the removal of artifactual desaturations from our statistical data analysis. This decreased the average time desaturated from 5.4% (220 minutes) to 4.2% (162 minutes) of the monitored period representing a 25% reduction in absolute incidence and a 35% reduction in episodic incidence of desaturation. Acquired data should be validated and inferences drawn from non-validated data must be assessed with caution.