Cradle Compression During Normothermic Machine Perfusion of Livers: Struggling Under the Weight of Expectation?

Normothermic machine perfusion (NMP) increases utilization of deceased liver grafts.1 It enables real-time assessment of graft viability and, by enabling grafts to be preserved for longer, offsets logistical problems such as concurrent liver offers and changes in the intended liver recipient. Cradle compression (CC) has emerged as a unique problem affecting liver grafts perfused on NMP. Essentially a computed tomography (CT) diagnosis, CC describes the heterogeneous opacification—typically in the dome of the right lobe—observed on imaging of recipients of liver grafts who had undergone NMP. During NMP, the liver is placed in a nonanatomical position with the vessels facing anteriorly, and CC is thought to reflect pressure necrosis and impaired portal inflow to the most dependent part of the liver. Martin et al2 have recently published the first series of CC following NMP, and report that radiological evidence of CC was present in 21 of 70 liver grafts (30%) imaged in the first 2 wk after transplant. Histopathological analysis of the CC lesions in 1 liver graft explanted early for hepatic artery thrombosis confirmed that the CT abnormalities corresponded with areas of diffuse contusion and necrosis of the involved hepatic parenchyma, but with preservation of the portal triads. Martin et al further proposed 3 categories of CC according to the severity of the lesions present in segments VII and VIII on CT imaging: a linear defect (grade 1), diffuse contusion (grade 2), or both a linear defect and diffuse contusion (grade 3). Predictive factors for the development of CC were heavy donor livers and livers from younger donors. One would perhaps anticipate that the presence of necrotic areas within the transplanted liver, no matter how circumscribed, would confer poorer transplant outcomes, but surprisingly, other than elevated early transaminase values, the presence of CC was not associated with poorer 1 y patient or graft survival. Thus, CC appeared to be a radiological entity without immediate clinical relevance. In this issue of Transplantation, Imai et al3 report that, in their center, CC was evident in 27 of 78 (34.6%) NMP liver grafts that were imaged in the first 2 wk after transplant. As with the study by Martin et al, a heavy donor liver (>1900 g) was an independent risk factor for the development of CC, but conversely, it occurred more frequently in livers from older donors. Critically, in the Imai series, the presence of CC was associated with poorer graft outcomes, with the incidence of early allograft dysfunction, acute kidney injury, and prolonged intensive care unit stay more common in recipients of liver grafts that developed CC. Postoperative bilirubin, international normalized ratio, and peak alanine aminotransferase levels were also higher in liver grafts with CC. Strikingly, whereas 90-d graft survival for liver transplants that either did not develop CC or developed only mild (grade 1) CC was 100%, this fell to 78% for livers with grade 3 CC. Of course, the association between the development of CC and poorer transplant outcomes does not necessarily imply a causative relationship, and it is possible that CC is instead a manifestation of a generally poorer quality liver graft. Against this, liver weight, as distinct from liver steatohepatosis, is not considered an independent risk factor for poor transplant outcome. One of the advantages of NMP is the ability to assess the viability of the liver using oxygen consumption and measurements of lactate, glucose, and transaminases within the circuit, which are used to inform transplant decisions.4 Neither the article by Imai et al nor by Martin et al provides details on the perfusion/biochemical characteristics of the livers while on the NMP circuit, and although presumably these parameters were within the acceptable range for transplantation, it is interesting to speculate whether these parameters were poorer for the livers that subsequently developed CC, or whether any particular perfusion characteristic predicted its development. Moving forward, a more detailed study of CC and its natural history is required. Notwithstanding, while awaiting these studies, it seems reasonable from the analysis by Imai et al to ascribe an independent deleterious impact of CC on liver transplant outcomes. Consequently, strategies aimed at ameliorating its development are warranted, particularly for heavier liver grafts. Some centers “massage” the liver graft while it is on the NMP circuit to try to augment portal perfusion by temporary elevation of the liver graft, although there are no clinical data to date that this reduces the incidence of CC. Other centers try to “flip” the liver on the machine to place it in its anatomical position with the vessels facing posteriorly, but these risks dislodging the cannulas, and again, there are as yet no data to support this maneuvre. An alternative approach is to incorporate a diaphragm into the NMP machine that evenly dissipates the pressure throughout the liver. None of the commonly used machines on the market (OrganOx metra, OCS Liver, TransMedics, Liver assist, XVIVO) has this feature, but the Liver4Life machine developed by Eshmuminov et al5 does use a diaphragm with the express aim of reducing pressure necrosis. They have demonstrated that with the inclusion of a diaphragm, the radioactive tracer FDG ([18F]-fluorodeoxyglucose) distributes homogenously throughout the liver while on their NMP machine. Whether this leads to a reduction in CC has, however, not yet been confirmed. In summary, with the ongoing expansion worldwide in the use of NMP, CC has emerged as a novel condition unique to livers placed on the circuit. Although its precise contribution to liver transplant outcomes is not yet clear and requires urgent further study, the development of strategies that attempt to prevent the development of CC is warranted.