Continuous, in vivo pulmonary venous admixture from fiberoptically measured hemoglobin saturations.

In six anesthetized swine, pulmonary venous admixture (Qsp/Qt) was calculated by four methods: a) Qsp/Qt 1, fiberoptically measured arterial and mixed venous Hgb saturation (SaO2 and SvO2), PaO2 and PvO2 derived from saturations; b) Qsp/Qt 2, fiberoptically measured SaO2 and SvO2, PaO2 and PvO2 measured by blood gas analysis; c) Qsp/Qt 3, PaO2 and PvO2 measured by blood gas analysis, SaO2 and SvO2 derived from tensions; d) Qsp/Qt 4, SaO2 and SvO2 measured by bench oximetry, PaO2 and PvO2 derived from saturations. Input from the fiberoptic catheters was fed into a computer programmed to calculate Qsp/Qt 1 every 20 sec. Fifty-eight of these values were compared with simultaneously calculated Qsp/Qt 2, 3, and 4. There was no difference between fiberoptic and derived SaO2 or fiberoptic and cooximetric SvO2. Correlations and slopes for Qsp/Qt 1 with Qsp/Qt 2, 3, and 4 were significant (p less than .05). Comparing mean differences, Qsp/Qt 1 was significantly different only from Qsp/Qt 3 (p less than .01). We conclude that dual oximetry reliably tracks Qsp/Qt.