The potential of accurate SvO2 monitoring during venovenous extracorporeal membrane oxygenation: an in vitro model using ultrasound dilution.

Introduction: Some degree of recirculation occurs during venovenous extracorporeal membrane oxygenation (VV ECMO) which, (1) reduces oxygen (O2) delivery, and (2) renders venous line oxygen saturation monitoring unreliable as an index of perfusion adequacy. Ultrasound dilution allows clinicians to rapidly monitor and quantify the percent of recirculation that is occurring during VV ECMO. The purpose of this paper is to test whether accurate patient mixed venous oxygen saturation (SvO2) can be calculated once recirculation is determined. It is hypothesized that it is possible to derive patient mixed venous saturations by integrating recirculation data with the ECMO circuit arterial and venous line oxygen saturation data.

Methods: A test system containing sheep blood adjusted to three venous saturations (low-30%, med-60%, high-80%) was interfaced via a mixing chamber with a standard VV ECMO circuit. Recirculation, arterial line and venous line oxygen saturations were measured and entered into a derived equation to calculate the mixed venous saturation. The resulting value was compared to the actual mixed venous saturation.

Results: Recirculation was held constant at 30.5 +/- 2.0% for all tests. A linear regression comparison of "actual" versus "calculated" mixed venous saturations produced a correlation coefficient of R2 = 0.88. Direct comparison of actual versus calculated saturations for all three test groups respectively are as follows; Low: 31.8 +/- 3.95% vs. 37.0 +/- 6.7% (NS), Med: 61.7 +/- 1.5% vs. 72.3 +/- 1.8% (p < 0.05), High: 84.4 +/- 0.9% vs. 91.2 +/- 1.1% (p < 0.05).

Discussion: There was a strong correlation between actual and calculated mixed venous saturations; however, significant differences between actual and calculated values where observed at the Med and High groups. While this data suggests that using quantified recirculation data to calculate SvO2 is promising, it appears that a straightforward derivative of the oxygen saturation-based equation may not be sufficient to produce clinically accurate calculations of actual mixed venous saturations.