Transesophageal versus surface pulse oximetry in intensive care unit patients

Objectives: To compare oximetric readings from the esophagus (STE0(2)) and the skin (finger, Sa0(2)) with those obtained from arterial blood samples ( Sa0(2)). In addition, to compare the influences of mean arterial pressure ( MAP) and body temperature to the accuracy of STE0(2) and Ss0(2) readings. Design: Prospective, single-center study. Setting: Surgical intensive care unit of an academic, teaching, and communi ty hospital in Austria. Patients: A total of 40 consecutive, severely traumatized or diseased, inte nsive care unit patients requiring mechanical ventilatory support and deep analgosedation. Patients had to be nonpregnant, greater than or equal to 19 yrs of age, and without a disease or a trauma of the esophagus. Interventions: Placement of an esophageal and a finger-pulse oximetry probe and a radial artery catheter. Measurements and Main Results: STE0(2), Ss0(2), MAP, and esophageal tempera ture were recorded continuously during a 4-hr period, and Sa0(2) was measur ed every 30 mins. The first outcome variable was the deviation of STE0(2) a nd Ss0(2) from Sa0(2), The second outcome variable was the influence of MAP and body temperature on STE0(2) and Ss0(2) regression analysis and repeate d measures. Analysis of variance was used for statistics (p < .05 was accep ted as significant), In patients with a MAP ranging from 29 to 111 mm Hg an d a temperature ranging from 33.4 degrees C (92.1 degrees F) to 39.2 degree s C (102,6 degrees F), Ss0(2) measurements underestimate Sa0(2) by 2% to 4% . Whereas STE0(2) matches Sa0(2), STE0(2) was not dependent on MAP or tempe rature, but increased temperature or low MAP were associated with falsely l ow Ss0(2) readings. Conclusions: Assuming correct positioning of the probe, readings from the e sophagus are more consistent with arterial oxygen saturation than readings from surface pulse oximetry, MAP or temperature changes do not influence ST E0(2) but they do affect Ss0(2). In critically ill patients, STE0(2), appea rs to be a more reliable variable than Ss0(2).