Single double-lumen venous–venous pump-driven extracorporeal lung membrane support

We sought to investigate the safety and feasibility of obtaining total respiratory support during 72 hours using a pump-driven (Levitronix CentriMag; Levitronix LLC, Waltham, Mass) venous–venous extracorporeal lung membrane (Novalung; Novalung GmbH, Hechingen, Germany) attached through a single double-lumen cannula (Novalung) into the femoral or jugular vein in pigs. Twelve pigs were initially mechanically ventilated for 2 hours (respiratory rate, 20–25 breaths/min; tidal volume, 10–12 mL/kg; fraction of inspired oxygen, 1.0; positive end-expiratory pressure, 5 cm H 2 O). Thereafter, the extracorporeal lung membrane was attached to the right femoral (n = 6, 26F) or jugular (n = 6, 22F) vein by using a single double-lumen cannula placed transcutaneously. Ventilatory settings were then reduced to near-apneic ventilation (respiratory rate, 4 breaths/min; tidal volume, 1–2 mL/kg; fraction of inspired oxygen, 0.21; positive end-expiratory pressure, 10 cm H 2 O), and pump flow was increased hourly until maximal efficacy. Blood gases and hemodynamics were measured hourly, and lung and plasma cytokine levels were measured every 4 hours. The device's mean blood flow was 2.16 ± 0.43 L/min, permitting an oxygen transfer and carbon dioxide removal of 203.6 ± 54.6 and 590.3 ± 23.3 mL/min, respectively. Despite static ventilation, all pigs showed optimal respiratory support, with a Pa o 2, Pa co 2, and mixed venous oxygen saturation of 226.2 ± 56.4, 59.7 ± 8.8, and 85.6 ± 5.3 mm Hg, respectively. There were no significant inflammatory, cellular, or coagulatory responses; lung cytokine levels remained in the normal range. Route (femoral vs jugular) or size (22F vs 26F) of the cannula did not change hemodynamic or respiratory parameters significantly. This circuit provides total respiratory support over 72 hours without inducing significant hemodynamic, coagulatory, cellular, or inflammatory responses.