Recirculation Quantified by Trans–Extracorporeal Membrane Oxygenation Thermodilution to Predict Mixed Venous Oxygenation during Veno-venous Extracorporeal Membrane Oxygenation: An In Vivo Porcine Study

BACKGROUND: Recirculation is a major contributor to refractory hypoxemia during veno-venous extracorporeal membrane oxygenation (V-V ECMO), yet it remains difficult to measure at the bedside. In an animal study, the authors applied the trans-extracorporeal membrane oxygenation thermodilution (TET) to quantify the phenomenon, identify its determinants, and estimate the mixed venous oxygenation noninvasively. METHODS: Eight pigs (65 ± 3 kg) were anesthetized and cannulated for femoral-jugular V-V ECMO. Acute respiratory distress syndrome was induced using oleic (n = 4) or hydrochloric acid (n = 4) targeting a Pa o2 /inspired oxygen fraction of 150 mmHg. The experiment lasted 24 h, with serial measurements of blood gases, hemodynamics, and respiratory mechanics. TET was performed by injecting a bolus of cold saline downstream the membrane lung and by recording the resulting temperature changes in the drainage and return cannulas. Recirculation was calculated as the ratio of the areas under the temperature-time curves. The determinants of the recirculation were analyzed with mixed linear models, including a random intercept. The measured recirculation was incorporated in a series of mass transfer equations to estimate the mixed venous oxygen content and saturation. RESULTS: Median recirculation was 7.4% [interquatile range, 1.1 to 22.0] and presented large intersubject variability. In 55% of measurements, recirculation was less than 10% (minimal recirculation). In the remaining measurements, the recirculation measured 27.4 [15.0 to 43.5; significant recirculation]. The ratio of extracorporeal blood flow to cardiac output measured 0.63 [0.55 to 0.72] and was the only physiologic factor correlated with recirculation (overall Pearson correlation coefficient, r = 0.67; in the significant recirculation, r = 0.81). The distance between the tip of the cannulas was, in our setup, nonsignificantly associated to the recirculation ( P = 0.74). An increased recirculation was modestly associated with a decreased mixed venous P o2 ( r = -0.37). The noninvasive estimation of the mixed venous oxygenation was highly accurate ( r = 0.98, bias plus 0.48 ml/100 ml for the oxygen content; r = 0.98, bias plus 4% for saturation). CONCLUSIONS: TET is a promising technique to measure the recirculation during V-V ECMO. Increasing the blood flow does not always lead to an improvement in oxygenation given the tight relationship between ratio of extracorporeal blood flow to cardiac output and recirculation. The proposed numerical method allows the noninvasive estimation of the mixed venous oxygenation without the need of a pulmonary artery catheter.