Modeling acid-base balance for in-series extracorporeal carbon dioxide removal and continuous venovenous hemofiltration devices.

Abstract

Patients with acute respiratory distress syndrome and acute kidney injury (AKI) treated by kidney replacement therapy may also require treatment with extracorporeal carbon dioxide removal (ECCO2 R) devices to permit protective or ultraprotective mechanical ventilation. We developed a mathematical model of acid-base balance during extracorporeal therapy using ECCO2 R and continuous venovenous hemofiltration (CVVH) devices applied in series for the treatment of mechanically ventilated AKI patients. Published data from clinical studies of mechanically ventilated AKI patients treated by CVVH at known infusion rates of substitution fluid without ECCO2 R were used to adjust the model parameters to fit plasma levels of arterial partial pressure of carbon dioxide (PaCO2 ), arterial plasma bicarbonate concentration ([HCO3 ]), and plasma pH (as well as certain other unmeasured physiological variables). The effects of applying ECCO2 R at an unchanged and a reduced tidal volume on PaCO2 , [HCO3 ] and plasma pH were then simulated assuming carbon dioxide removal rates from the ECCO2 R device measured in the clinical studies. Agreement of such model predictions with clinical data was good whether the ECCO2 R device was positioned proximal or distal to the CVVH device in the extracorporeal circuit. Although carbon dioxide removal rates from the ECCO2 R device measured in one previous clinical study were higher when it was placed proximal to the CVVH device, suggesting that such in-series positioning was optimal, the current mathematical model demonstrates that proximal positioning of the ECCO2 R device also results in lower bicarbonate (and, therefore, total carbon dioxide) removal from the distal CVVH device. Thus, the removal of total carbon dioxide by such extracorporeal circuits is relatively independent of the position of the in-series devices. It is concluded that the described mathematical model has quantitative accuracy; these results suggest that the overall acid-base balance when using ECCO2 R and CVVH devices in a single extracorporeal circuit will be similar, independent of their in-series position.