D-dimer Release From Livers During Ex Situ Normothermic Perfusion and After In Situ Normothermic Regional Perfusion: Evidence for Occult Fibrin Burden Associated With Adverse Transplant Outcomes and Cholangiopathy

Abstract

Background: Deceased donor livers are prone to biliary complications, which may necessitate retransplantation, and we, and others, have suggested that these complications are because of peribiliary vascular fibrin microthrombi. We sought to determine the prevalence and consequence of occult fibrin within deceased donor livers undergoing normothermic ex situ perfusion (NESLiP) and evaluate a role for fibrinolysis.

Methods: D-dimer concentrations, products of fibrin degradation, were assayed in the perfusate of 163 livers taken after 2 h of NESLiP, including 91 that were transplanted. These were related to posttransplant outcomes. Five different fibrinolytic protocols during NESLiP using alteplase were evaluated, and the transplant outcomes of these alteplase-treated livers were reviewed.

Results: Perfusate D-dimer concentrations were lowest in livers recovered using in situ normothermic regional perfusion and highest in alteplase-treated livers. D-dimer release from donation after brain death livers was significantly correlated with the duration of cold ischemia. In non-alteplase-treated livers, Cox proportional hazards regression analysis showed that D-dimer levels were associated with transplant survival ( P = 0.005). Treatment with alteplase and fresh frozen plasma during NESLiP was associated with significantly more D-dimer release into the perfusate and was not associated with excess bleeding postimplantation; 8 of the 9 treated livers were free of cholangiopathy, whereas the ninth had a proximal duct stricture.

Conclusions: Fibrin is present in many livers during cold storage and is associated with poor posttransplant outcomes. The amount of D-dimer released after fibrinolytic treatment indicates a significant occult fibrin burden and suggests that fibrinolytic therapy during NESLiP may be a promising therapeutic intervention.

Graphical abstract

FIGURE 1.

Cambridge protocol for using in situ normothermic regional perfusion and ex situ normothermic liver perfusion. DCD livers were preferentially recovered using NRP, but where that was not possible, DCD livers underwent NESLiP. DBD livers underwent NESLiP if there was doubt about viability or where a long preservation time was envisaged for recipient or logistical reasons. DBD, donation after brain death; DCD, donation after circulatory determined death; NESLiP, normothermic ex situ perfusion; NRP, in situ normothermic regional perfusion.

FIGURE 2.

Flow diagram illustrating perfusates studied by donor type and treatment. Samples of perfusate from 163 livers taken after 2 h of NESLiP were analyzed. These represent samples from all the liver perfusions in which ethical approval had been obtained for research where samples existed. DBD, donation after brain death; DCD, donation after circulatory determined death; NESLiP, normothermic ex situ perfusion; NRP, in situ normothermic regional perfusion.

FIGURE 3.

Box and whisker plot of perfusate D-dimer concentrations after 2 h of NESLiP by donor type and treatment. The box represents the interquartile range, the horizontal line the median of the group, and the whiskers the range. There was a significant difference in 2-h D-dimer levels across all groups (Kruskal-Wallis P < 0.001), with a difference also when the alteplase-treated DCD livers are ignored (Kruskal-Wallis P < 0.001). Statistical significance between groups was estimated by the Kolmogorov-Smirnov test and is indicated on the chart. The alteplase-treated group also had D-dimer levels that were significantly higher than NRP (P < 0.001), DCD livers (P = 0.008), and DBD livers (P = 0.041). DBD, donation after brain death; DCD, donation after circulatory determined death; NESLiP, normothermic ex situ perfusion; NRP, in situ normothermic regional perfusion.

FIGURE 4.

D-dimer concentrations in the preperfusion effluent of the livers. Livers were flushed with 2 L of Hartmannn’s solution before being placed on the perfusion machine, and the effluent from the vena cava was collected and sampled. Seventeen of the 20 livers studied had effluent concentrations of D-dimers over 100 ng/mL. There were 3 DCD livers with particularly high D-dimer concentrations: liver A failed viability testing, liver B suffered primary nonfunction, and liver C developed severe vasoplegia requiring epinephrine and external cardiac massage. DBD, donation after brain death; DCD, donation after circulatory determined death.

FIGURE 5.

Transplant survival according to 2-h perfusate D-dimer concentrations. Transplant survival (graft survival not censored for death) is worse with increasing D-dimer concentration quartile.

FIGURE 6.

Cumulative incidence of nonanastomotic stricture by 2-h D-dimer quartile. Increasing D-dimer concentration quartiles were associated with an increased incidence of nonanastomotic biliary strictures.

FIGURE 7.

Rate of change of D-dimer concentrations over time with and without alteplase/FFP treatment. D-dimer concentrations (median) in the perfusate of 19 livers that underwent NESLiP and that had received either bolus or an infusion of alteplase with FFP and 8 livers that did not receive alteplase and FFP. FFP, fresh frozen plasma; NESLiP, normothermic ex situ liver perfusion.

FIGURE 8.

Intraoperative blood transfusion requirements of alteplase-treated and comparator livers. A, Number of packed red cell units given. This shows the number of units of packed red cells given before and after implantation in each case for livers treated with alteplase (n = 9) or not receiving alteplase (n = 18) during NESLiP. B, Proportion of packed red cell units given postimplantation as proportion of total blood units used. This figure shows the same data as above, but the number of units given postimplantation is given as a percentage of the total blood replacement during surgery. NESLiP, normothermic ex situ perfusion.