Accuracy and clinical utility of an oxygen saturation catheter.

Twenty-three fiberoptic catheter oximeters were placed in the umbilical arteries of 22 neonates to determine the accuracy and reliability of the sensor and clinical utility of continuous measurement of arterial oxygen saturation. Comparison of fiberoptic saturation readings with 1039 bench measured blood samples revealed a correlation coefficient of 0.97 with a mean error of 0.74 +/- 2.17 (SD) saturation. Although 11 catheters remained reasonably accurate (in error less than 3% saturation) for up to 488 h of use, 12 required in vivo calibration at a median time of 31 h after initial insertion. Clinical use of the catheter oximeter was evaluated in 11 infants by comparison to matched controls. Control infants required 55% more arterial blood samples for care, twice the number of transfusions, and spent twice the amount of time with a PaO2 greater than 110 torr. The catheter oximeter was found to be a useful clinical tool. Saturation values up to 95% reflected accurately and reliably an infant's arterial saturation status, providing constant and comprehensive assessment of clinical variations and the effects of therapy. The most useful feature was its rapid and accurate response to hypoxic incidents and to efforts to correct these conditions for improved care of infants in life-threatening situations.