Air embolism as a complication of venovenous bypass during liver transplant for diffuse hemangiomatosis

A 48-year-old female patient, who was 162 cm tall and weighed 102 kg, underwent a first liver transplant for diffuse hemangiomatosis. Preoperatively, she displayed a restrictive syndrome on respiratory function tests and increased dimensions of the right atrium and right ventricle on an echocardiogram. Renal function was normal and laboratory tests showed 10.3 g/L hemoglobin, prothrombin time of 5.0 seconds above control, and normal values of platelet count, albumin, liver enzymes, and bilirubin. On arrival at the operating room, the arterial pressure was 133/63 mm Hg and heart rate 85 beats/min. Anesthetic induction was performed with thiopental, fentanyl, and succinylcholine; anesthesia maintenance utilized isoflurane vaporized in oxygen/air, fentanyl, and vecuronium. Monitoring included invasive arterial pressures, pulmonary artery catheter, pulse oximetry, and a Vigilance monitor (Edward-Baxter, Irvine, Calif, USA) for continuous evaluation of cardiac output and pulmonary artery oxygen saturation. The native liver displayed remarkable dimensions and a strong tendency to laceration despite careful technique. For vascular decompression of the surgical field, a porto-femoro-axillary venous bypass was initiated with a vortex Biomedicus pump (Minn, USA), using transparent heparin-coated tubes and a disposable head pump (Jostra AG; Hirrlingen, Germany, reference BE-RF-32.) An air embolism was detected a few minutes after the initiation of bypass by decreases in the end-tidal partial pressure of carbon dioxide (PetCO2) from 26 to 16 mm Hg and in the arterial saturation (SaO2) from 97% to 93%, with increased pulmonary artery pressures (PAP) from 34/19/26 to 57/26/39 mm Hg, a slight decrease in systemic arterial pressure (AP) from 114/65 to 102/75 mm Hg, and an increased heart rate from 94 to 101 beats/min. Neither careful visual inspection nor the use of air detectors in the equipment were able to detect air inside the bypass system, but the femoral cannula showed a small hole that occurred due to a traumatic clamp used for cannula insertion. The hole was totally covered with an adhesive plastic band and the signs of air embolism disappeared within a few minutes. While still on bypass, 2 hours and 50 minutes later, the AP decreased over 10 minutes from 112/65 to 65/43 mm Hg with increased PAP from 26/12/18 to 68/38/48 mm Hg and central venous pressure (CVP) from 9 to 18 mm Hg. PetCO2 decreased from 27 to 12 mm Hg, SaO2 from 99% to 87%, and heart rate from 112 to 87 beats/min (Fig 1). The adhesive band that covered the hole of the femoral cannula had become wet, lost its adherence, and allowed entry of air again, once more not detected by the equipment detectors or by visual inspection. Bypass was immediately stopped; the patient was treated with 50 g of epinephrine, 0.5 mg of atropine, and 100% oxygen, recovering acceptable pressures in 12 minutes. Surgery continued and graft reperfusion was achieved 17 minutes after detection of this second episode of air embolism. No more hazardous events were observed; the patient was extubated some hours later. The further evolution was normal and no sequelae were observed after this occurrence. In our overall experience of bypass used in 125 (27%) of 462 transplants, this is the sole case of air embolism with an origin in the bypass, leading to an incidence of 0.8% for this complication.