Conjunctival oxygen tension and its relationship to arterial oxygen tension.

Using a miniaturized Clark electrode embedded in a polymethylmethacrylate eyepiece, we measured transconjunctival oxygen tension (PcjO2) in 5 healthy volunteer subjects at multiple levels of steady-state isocapnic hypoxia, normoxia, and hyperoxia. PcjO2 was linearly related to arterial oxygen tension (PaO2) as PaO2 ranged from 28 to 205 mm Hg (PcjO2 = 0.59 PaO2 + 0.36 mm Hg; r = 0.94; standard error of the estimate = 7.09 mm Hg). However, the relationships between PcjO2 and PaO2 varied significantly among subjects. Whereas the overall mean ratio of PcjO2 to PaO2 was 0.59, the mean ratio for subjects ranged from 0.47 to 0.79 and was significantly different among subjects (P less than 0.0001). The time response of the electrode to a step change in oxygen tension in vitro was exponential, with a 90% response time of 38 seconds after a lag of 3.7 seconds. The time responses to in vivo changes in oxygen tension were also exponential. From hypoxia to normoxia, 90% response time was 45.0 seconds after a lag of 5.1 seconds; from room air to hypoxia, 90% response time was 72.4 seconds after a lag of 30.3 seconds; from room air to hyperoxia, 90% response time was 87.2 seconds after a lag of 6.8 seconds. We conclude that, although PcjO2 measured by a miniaturized Clark electrode is linearly related to PaO2 in healthy subjects, variation in the relationship of PcjO2 to PaO2 among individuals will prevent precise estimation of PaO2 for any individual unless subject-specific calibration is performed.