The accuracy and response characteristics of a simplified ear oximeter.

We evaluated the accuracy and speed of response of a newly available, lightweight, and relatively inexpensive ear oximeter (Biox II oximeter). The instrument was compared with another oximeter (Hewlett-Packard 47201A) under conditions of both steadily maintained and progressively increasing hypoxia induced in normal subjects by rebreathing. The new oximeter (Biox II), which can be operated in either a "normal" or "fast" response mode, as selected by a switch on the front panel, was evaluated in its "normal" mode during steady-state hypoxic conditions and in both "normal" and "fast" modes during progressive hypoxic conditions. The other oximeter (HP 47201A) was operated in its factory preset "normal" mode for all measurements. During steady-state hypoxia the relationship between oximetric arterial oxygen saturation (SaO2) readings (y) and spectrophotometrically measured SaO2 in samples of arterial blood (x) when SaO2 exceeded 65 percent was as follows: for the new oximeter (Biox), y = 0.95x + 3.25 (r = 0.96); and for the other oximeter (HP 47201A), y = 1.03x - 2.31 (r = 0.94). Neither of these relationships differed significantly from the line of identity. During trials of progressive isocapnic hypoxia induced acutely in ten normal subjects, SaO2 was measured continuously by both oximeters. With the new oximeter (Biox) operated in the "normal" mode, the relationship between values for SaO2 from it (y) and the other oximeter (Hewlett-Packard) (x) was y = 0.85x + 12.91 (r = 0.93). When the new oximeter (Biox) was switched to its "fast" response mode, the relationship more closely approximated the line of identity such that y = 1.05x - 5.95 (r = 0.98). The response of the new oximeter (Biox II) to an in vitro step change in saturation followed a complex nonexponential function characterized by small initial changes in output signal with the greatest changes in output occurring during the latter portion of the response period. The 50 percent response times of the new oximeter (Biox II) were 5.65 seconds and 2.86 seconds in the "normal" and "fast" modes, respectively, by contrast to the 50 percent response time of 2.87 seconds for the other oximeter (H-P 47201A). We conclude that the new oximeter (Biox II) demonstrated accuracy comparable to a more complex and expensive oximeter and had response characteristics that may be useful in clinical and laboratory settings.