A New Era of Decision-making in Liver Transplantation

Objectives: Flavin mononucleotide (FMN), a marker of mitochondrial complex 1 injury, has not yet been validated for its predictive value of outcomes and economic impact. Background: Normothermic machine perfusion (NMP) is the only ex situ perfusion technique currently approved for liver transplantation in the United States. Optimal graft viability assessment on this approach remains controversial. Methods: All liver transplants at our center were included, divided into static-cold storage (n=418), NMP (OrganOx Metra ) with traditional viability criteria (October 2022 to January 2024, n=213), and prospective viability assessment using FMN (NMP+FMN, January 2024 to August 2024, n=143). Perfusate fluorescence spectroscopy was performed to quantify FMN during NMP. Spectroscopy results were correlated with tissue analyses. Standard risk factors and clinically relevant core outcomes were collected for analysis. Groups were propensity-matched, and posttransplant outcomes, including economics, were assessed using inverse probability of treatment weighting. Mixed-effects models assessed complications, graft loss, and FMN-guided liver utilization. A decision-analytic model was used to assess the cost-benefit of NMP and FMN testing. Results: Graft loss was predicted best by perfusate FMN (>1700 samples; c-statistic AUC 0–4 h NMP: 0.96, 95% CI: 0.93–0.97, P <0.0001) versus traditional viability markers. High FMN grafts demonstrated significantly more mitochondrial injury measured in tissues at the end of NMP. Since implemented prospectively, FMN-based viability assessment during NMP led to a comparable liver utilization rate of NMP=94 versus NMP+FMN=90% ( P =0.346) despite higher overall donor and recipient risk. Over one third (n=43, 35%) were livers from donation after circulatory death donors (DCD). Elevated perfusate FMN of >1.75 μg/mL at 4 hours was independently associated with reduced graft survival and death-censored graft survival. Liver transplants in the FMN-era were independently associated with improved graft survival on Cox regression (HR: 6.841, 95% CI: 1.447–37.300, P <0.001). Risk-adjusted outcomes, including biliary and overall complications, major (Clavien>IIIA) complications, liver-related major complications, and graft loss, were improved with FMN-based viability testing. Overall morbidity measured by the comprehensive complications index (CCI) was reduced with NMP but did significantly decrease with additional FMN use compared with static-cold storage. Such results were upheld when DBD and DCD grafts were evaluated independently. Liver transplantations with high FMN livers demonstrated greater cumulative costs ( P <0.001). On mixed-effects modeling, 44% percent of transplant-related cost variation was explained by FMN in the top quintile (>1.75 μg/mL). Risk-matched FMN-tested DBD grafts specifically demonstrated an incremental 16% reduction in major complications with a net $33,657 saving per graft in the decision-analytic model, while DCD grafts demonstrated 30% improvement in major complications and an incremental cost-reduction of $53,563 per graft. Conclusions: Our findings support routine utilization of FMN-based viability assessment during NMP. Despite higher donor/recipient risk, our center has reduced complications and improved graft survival with FMN-based decision-making. Reduced transplant costs likely stem from a reduction in posttransplant complications.