Background: Flavin mononucleotide (FMN) is released from complex I during anoxia, it reflects directly on mitochondrial preservation damage and the remaining metabolic capability of the cell. Its release into the perfusate at perfusion start is the base of its reliable prediction of graft dysfunction. FMN is currently the only option for organ assessment during hypothermic oxygenated machine perfusion (HOPE).
Methods: During 50 HOPE, FMN was measured and correlated with standard laboratory and follow-up parameters including complications according to Dindo et al., early allograft dysfunction (EAD) defined by Olthoff et al., biliary complications, 1-year graft and patient survival. FMN was measured by fluorescence spectroscopy, levels were displayed in artificial units and concentrations determined with generation of a standard curve.   
Results: The area under the curve (AUC) for prediction of EAD at 5 minutes of perfusion was 0.744, with the identified cut-off of 10.65 ng/dL (sensitivity 87%, specificity 63%), 52% of recipients that developed an EAD were identified. FMN levels at 5 minutes were higher in grafts whose recipients did not survive the first year after liver transplantation (p<0.001**). Patient 1-year survival was predicted at 5 minutes with an AUC of 0.890. The determined cut-off of 23.5ng/dL (sensitivity 83.3 % and specificity 92.7 %) was used to identify patients with a high risk for early mortality. Below the cut-off, 98% of recipients survived whereas only 37.5% of recipients above the cut-off survived. Both cut-offs combined allowed for risk stratification in high (>23.5 ng/dL), intermediate (between 10.65 and 23.5 ng/dL) and low risk (<10.65 ng/dL), rate of EAD development (p=0.004), 1-year mortality (p<0.001**), need for dialysis (0.019) and re-transplantation (p=0.001) was highest in the high-risk group followed by the intermediate and low risk group.
Conclusion: FMN allows for excellent risk stratification in liver grafts during HOPE. Categorization in risk groups helps to identify grafts with an especially high risk for graft loss and early mortality. The threshold identified in this study enables reliable prediction of patient survival as early as 5 minutes after perfusion start. In addition, FMN levels correlate to increasing rates of complications and higher morbidity underlining the relevance and potential of FMN in hypothermic perfusion viability testing.