Intravenacaval membrane oxygenation and carbon dioxide removal in severe acute respiratory failure.

STUDY OBJECTIVE

To characterize the physiologic response to, and safety of, intravenacaval membrane oxygenation and carbon dioxide removal.

DESIGN

Interventional before-after study.

SETTING

University teaching hospital ICU.

PATIENTS

Twenty-two patients with severe acute respiratory distress syndrome (ARDS).

INTERVENTIONS

Implantation of a hollow-fiber membrane oxygenator (IVOX; CardioPulmonics; Salt Lake City, Utah) into the superior and inferior venae cavae by venotomy of the right femoral or right internal jugular vein for a duration of up to 20 days.

MEASUREMENTS

Hemodynamic measurements using pulmonary artery and systemic artery catheters, ventilator settings (FIO2, minute ventilation, peak inspiratory pressure, and positive end-expiratory pressure), arterial and mixed venous blood gases (pH, PCO2, PO2, and measured saturation), and clinical laboratory determinations (CBC, fibrinogen, plasma hemoglobin, complement C3 and C5) were obtained. Calculations of PaO2/FIO2 ratio and PaCO2-VE product were used to assess gas exchange efficacy. Microbiologic cultures were obtained from the device and wound following explantation. Survival to ICU discharge and hospital discharge were recorded.

RESULTS

Implantation was successful in 20 of 22 patients. Gas exchange rates averaged 50.4 +/- 15.8 mL.min-1 for carbon dioxide and 71.1 +/- 20.2 mL.min-1 for oxygen. A reduction in FIO2 from 0.78 +/- 0.16 to 0.63 +/- 0.21 and in VE from 177 +/- 94 mL.kg-1.min-1 to 127 +/- 58 mL.kg-1.min-1 was possible within 4 h post-implantation. By 12 h, FIO2 was reduced to 0.57 +/- 0.18. Indices of gas exchange improved significantly after implantation, with PaO2/FIO2 ratio increasing from 79 +/- 20 to 112 +/- 47 and PaCO2-VE product decreasing from 7.6 +/- 4.2 to 4.9 +/- 2.5 within 4 h. A significant reduction in peak inspiratory pressure was achieved (45 +/- 10 to 38 +/- 9 cm H2O). Major complications were blood loss during implantation requiring transfusion in 11 patients, a retroperitoneal bleed in 1 patient, and femoral deep venous thrombosis in 4 patients, but there were no long-term sequelae or IVOX-related deaths. The ICU and hospital survival were 10/20 (50%) and 8/20 (40%), respectively.

CONCLUSIONS

Intravenacaval membrane oxygen and carbon dioxide removal can provide partial respiratory support during severe respiratory failure and permit reductions in the level of mechanical ventilator support, with an acceptable safety profile.