Predictive value of early postoperative lactate (<6 h) during normothermic machine perfusion and outcome after liver transplantation: results from a multicentre study

Abstract

Background: Biomarkers with strong predictive capacity towards transplantation outcome for livers undergoing normothermic machine perfusion (NMP) are needed. We investigated lactate clearing capacity as a basic function of liver viability during the first 6 h of NMP.

Methods: A trial conducted in 6 high-volume transplant centres in Europe. All centres applied a back-to-base NMP approach with the OrganOx metra system. Perfusate lactate levels at start, 1, 2, 4 and 6 h of NMP were assessed individually and as area under the curve (AUC) and correlated with EAD (early allograft dysfunction), MEAF (model for early allograft function) and modified L-GrAFT (liver graft assessment following transplantation) scores.

Results: A total of 509 livers underwent ≥6 h of NMP before transplantation in 6 centres in the UK, Germany and Austria. The donor age was 53 (40-63) years (median, i.q.r.).The total NMP time was 10.8 (7.9-15.7) h. EAD occurred in 26%, MEAF was 4.72 (3.54-6.05) and L-GrAFT10 -0.96 (-1.52--0.32). Lactate at 1, 2 and 6 h correlated with increasing robustness with MEAF. Rather than a binary assessment with a cut-off value at 2 h, the actual 2 h lactate level correlated with the MEAF (P = 0.0306 versus P = 0.0002, Pearson r = 0.01087 versus r = 0.1734). The absolute lactate concentration at 6 h, the AUC of 0-6 h and 1-6 h (P < 0.0001, r = 0.3176) were the strongest predictors of MEAF.

Conclusion: Lactate measured 1-6 h and lactate levels at 6 h correlate strongly with risk of liver allograft dysfunction upon transplantation. The robustness of predicting MEAF by lactate increases with perfusion duration. Monitoring lactate levels should be extended to at least 6 h of NMP routinely to improve clinical outcome.

Graphical Abstract

Fig. 1

Perfusate lactate levels during normothermic machine perfusion (NMP) for the surviving recipients (solid boxes) and non-surviving (open triangles) (a) and for donation after brain death (solid circles) or donation after circulatory death (open boxes) (b) as mean ± s.d c Linear regression analysis of perfusate lactate as a function of cold ischaemia time after start of NMP.

Fig. 2

Area under the curve (AUC) calculations AUC was determined with incorporation of the lactate measurements after the start of normothermic machine perfusion (NMP), 1 and 2 h (a, AUC 0–1–2), additionally with 4 h (b, AUC 0–1–2–4) or 4 and 6 h (c, AUC 0–1–2–4–6). AUC was determined with omitting the first lactate measurement after the start of NMP with incorporation of 1 and 2 h measurements (d, AUC 1–2), additionally with 4 h (e, AUC 1–2–4) or 4 and 6 h (f, AUC 1–2–4–6).

Fig. 3

Linear regression analysis against model for early allograft function (MEAF) and the single time point measurements (a–e) and calculated AUCs (f–k) AUC, area under the curve

Fig. 4

Forest plot of mixed model for repeated measurements Symbols show the effect size along with 95% confidence intervals.