A pragmatic alternative to commercial perfusion devices: routine liver machine perfusion using a heart–lung machine

Introduction: Machine perfusion (MP) is one of the most effective preservation strategies for suboptimal liver allografts but remains limited by the high cost and restricted availability of commercial perfusion devices in different countries and transplant centers. Adaptation of equipment for the cardiopulmonary bypass may enable broader implementation of MP.
Methods: A perfusion circuit was developed for the main MP types by modifying a standard disposable for cardiopulmonary bypass (CPB). A heart–lung machine (HLM) was used as the perfusion device. This setup is routinely applied in our center from the beginning of the MP program due to the lack of access to commercial liver perfusion devices. The system also allows integration of continuous renal replacement therapy into the circuit.
MP was performed using standardized protocol. All grafts were procured from extended criteria brain-dead donors. Data from all DHOPE-COR and DHOPE-COR-NMP procedures, with or without ischemia-free or hypoxia-free implantation, performed between July 2024 and March 2026 (n = 90) were collected and analyzed.
Results: DHOPE-COR was performed in 74,4% of cases (n=67, including hypoxia-free liver implantation (n=9)). Median perfusion time was 160,00 [IQR 103,00; 223,50] min, and median cold ischemia time was 321,00 [IQR 216,75; 412,75] min. EAD (Olthoff criteria) was observed in 10,5% and PNF rate was in 1,5%. 90-day patient mortality and 90-day graft loss were 8,5% and 2,7%, respectively.
DHOPE-COR-NMP was performed in 25,6% of cases (n=23, including ischemia-free liver implantation (n=2)). Median perfusion time was 621,00 [IQR 459,00; 706,50] min, median cold ischemia time - 242,00 [IQR 215,00; 297,00] min. EAD was observed in 73,3%, mostly due to high enzymes level (90,9%). No PNF or retransplantation were observed in this cohort. Median donor macrosteatosis was 75% [IQR 30-90%]. 90-day graft loss and patient mortality were 0%. Overall graft utilization rate was 65,2%. No device failures were observed.
Conclusion: Heart–lung machine–based MP represents a safe and accessible approach using equipment and disposables commonly available in centers with cardiothoracic surgery. This strategy may facilitate wider adoption of MP, particularly in regions with limited access to commercial devices; however, the main limitation, in our opinion, is the need for the constant monitoring by the surgeon as the system is flow-controlled rather than pressure-controlled and is not designed for autonomous work.