Retrospective comparison of equine hemoglobin oxygen saturation measured by a human-specific co-oximeter, or derived from an algorithm using temperature-corrected and -uncorrected oxygen tension

OBJECTIVE

Hemoglobin oxygen saturation (SO₂) is measured by co-oximetry (SO_2cox) or calculated from partial pressure of oxygen (PO₂) using algorithms (SO_2alg). To compare venous and systemic arterial blood sample data retrospectively and to examine whether temperature correction of PO₂ is important.

STUDY DESIGN

Retrospective study.

ANIMALS

A group of 21 healthy, adult, sedated or anesthetized horses.

METHODS

Mixed-venous and systemic arterial blood samples (1 mL) were anaerobically collected using commercial preheparinized syringes from the right ventricle and facial artery, respectively. Blood was analyzed using a commercial gas analyzer and human-specific co-oximeter within 10 minutes of collection or stored on ice and analyzed within 30 minutes. PO₂ was measured at 37 °C and corrected using body temperature (T_core) from a pulmonary artery catheter thermistor. SO_2cox and hemoglobin subtypes were measured by co-oximetry (37 °C). An algorithm developed for Thoroughbred horse blood was used to calculate SO_2alg using PO₂ at 37 °C and SO_2algcorr with PO₂ corrected to T_core. SO_2alg and SO_2algcorr were each paired with SO_2cox using Bland-Altman (repeated measures) ratio of SO_2alg/SO_2cox (204 samples).

RESULTS

SO_2alg overestimated SO_2cox when PO₂ was <80 mmHg (10.7 kPa); ratio and limits of agreement: 1.2 (0.9-1.6) but became accurate when PO₂ was ≥80 mmHg: 1.0 (1.0-1.0). With all data, SO_2algcorr did not differ from SO_2alg:1.1 (0.8-1.4). Methemoglobin (FMetHb) and carboxyhemoglobin (FCOHb) were significantly higher in venous [FMetHb: median (range): 1.8 (0-2.9)%; FCOHb: 0.1 (0-2)%] than in arterial blood [FMetHb: 0.5 (0-2.2)%; FCOHb: 0 (0-0.3)%].

CONCLUSIONS

The algorithm appeared robust when PO₂ was ≥80 mmHg (10.7 kPa) but overestimated when PO₂ was lower. Temperature correction was not important within 34.9-37.3 °C.

CLINICAL RELEVANCE

SO_2alg overestimation in venous blood can result in calculation of higher intrapulmonary shunt fraction than SO_2cox values.