All consecutive liver transplantations (LTs) performed at Paul Brousse University Hospital (Villejuif, France) between April 2022 and April 2023 were included in the study. Inclusion criteria were patients aged ≥18 years with end-stage liver disease. Exclusion criteria included LT from living donors, split liver grafts, re-transplantation, LT for fulminant hepatitis, use of venovenous bypass during LT, and lack of consent for the use of clinical data. This prospective observational exploratory study was approved by the local Ethics Committee (CO 16-006; CNIL No. 1856085).

After procurement, liver grafts were preserved in Institut Georges Lopez-1 (IGL-1®) solution at 4°C until transplantation. Before implantation, all grafts were flushed with 800 mL of 5% albumin solution, according to the protocol of our institution. Total ischemia time was defined as the interval between cessation of blood supply at the time of liver graft procurement and portal reperfusion upon implantation in the recipient. The balance of risk (BAR) score is a composite measure that includes the MELD score, recipient age, donor type (living donor), emergency status of LT, re-transplantation status, and cold ischemia time. It is used to estimate prognosis and survival in liver transplant recipients. A BAR score <9 predicts an excellent prognosis, whereas a score ≥18 indicates a poor prognosis with increased mortality at 3 months post-LT (20). An extended criteria donor (ECD) was defined as a liver graft with macrovacuolar steatosis ≥30%, donor age ≥70 years, elevated GGT, or from a donation after circulatory death (DCD) donor. No liver grafts were perfused using oxygenated dynamic perfusion during the study period.

The retrohepatic inferior vena cava was preserved whenever possible. At our institution, a temporary porto-caval shunt is systematically performed during liver explantation to prevent mesenteric congestion. During graft implantation, a vena cava piggyback anastomosis was created on the hepatic veins. Other anastomoses, including porto-portal, arterial, and biliary or bilio-jejunal, were performed based on the usual practice of the surgical team. Vascular unclamping (caval and portal) was performed simultaneously, followed by arterial unclamping. Abdominal drainage was systematic. Immunosuppression was initiated at the time of surgical incision with methylprednisolone 5 mg/kg (maximum 500 mg). Anesthetic management was at the discretion of the anesthesiologist.

The objective was to assess glycocalyx alteration during liver graft procurement and liver transplantation as a marker of ischemia–reperfusion injury and to determine whether glycocalyx alteration is predictive of early liver graft dysfunction.

The primary endpoint was early allograft dysfunction (EAD), which was evaluated based on the MEAF score. The MEAF is a continuous score calculated from ALT, INR, and bilirubin levels measured on postoperative day 3. It is calculated as follows: MEAF = score ALT_max3POD + score INR_max3POD + score bilirubine_3POD, where score ALT_max3POD = 3.29/(1 + e^−1.9132(ln (ALT_max3POD)) − 6.1723); score INR_max3POD = 3.29/(1 + e^−6.8204(ln (score INR_max3POD)) − 0.6658); score bilirubine_3POD = 3.4/ (1 + e^−1.8005(ln(bilirubine_3POD)) − 1.0607). Severe graft dysfunction is defined as a MEAF score ≥9, which corresponds to a 30% risk of graft loss and a 40% risk of patient death within 3 months (1).

The secondary end points were post-LT outcomes: (1) primary non-function corresponding to graft non-function within 7 days post-LT without vascular cause, leading to death or urgent re-transplantation (21); (2) severe post-LT complications defined according to Dindo–Clavien classification as type III, which include severe complications requiring endoscopic, radiological, or surgical intervention (22).

During liver graft procurement, the liver was flushed with IGL-1 solution after the aorta was clamped, ensuring liver graft preservation without blood. To assess endothelial glycocalyx degradation and its excretion kinetics, samples of liver vena cava effluent (blood or preservation fluid/albumin) were collected at specific time points, listed as follows: (i) during liver graft procurement before aortic clamping, (ii) at the end of cold ischemia preservation, after the graft had been flushed with 5% albumin solution through the portal vein, and (iii) 5 min after liver graft reperfusion in the recipient, i.e., after caval and portal unclamping.

Blood samples were collected in citrated or EDTA tubes. Blood was centrifuged at 2,000g for 10 min to obtain plasma. Plasma samples were aliquoted in 0.5 mL Eppendorf tubes and stored at −80 °C until further analysis by enzyme-linked immunosorbent assay (ELISA). To assess endothelial glycocalyx alteration, we quantified the following biomarkers in the collected samples using ELISA assay kits: Syndecan-1 (human CD138; SDC1, #EHSDC1, Thermo Fisher Scientific, USA), heparan sulfate (E-EL-H2364c, Elabscience Biotechnology Co., France), angiopoietin-1 (EPX010-12235-901, Thermo Fisher Scientific), and angiopoietin-2 (KHC1641, Thermo Fisher Scientific). Ischemia–reperfusion injury was evaluated using inflammation and cell damage markers. Inflammation was assessed by quantifying TNF-alpha (BMS223-4, Thermo Fisher Scientific). Cell death was evaluated by quantifying total cytokeratin 18 (CK-18), a fragment released into circulation upon cell death, reflecting the overall cell death (M65 EpiDeath ELISA, PEVIVA, Eurobio Scientific, France). Cleaved cytokeratin 18 (c-CK-18) is a component generated by enzymatic digestion during apoptosis, enabling the assessment of programmed cell death (M30 Apoptosense Chronic Liver Human, PEVIVA, Eurobio Scientific). Necrosis was indirectly evaluated by subtracting cleaved CK-18 from total CK-18.

To evaluate ischemia–reperfusion injury, surgical liver biopsies were performed (1) before aortic clamping in the donor and (2) at the end of LT in the recipient (before abdominal closure). Hepatic tissue damage was assessed by a pathologist, who evaluated neutrophil infiltration, hepatocyte ballooning, and necrosis and classified the ischemia–reperfusion injury as severe, moderate, mild, or absent. We determined whether there was a correlation between histological lesions and excretion of synd-1 or heparan sulfate during liver graft procurement. Histological evaluation was performed blinded to glycocalyx measurements.

As part of routine clinical care, all patients underwent daily blood tests. Clinical data were collected before, during, and up to 90 days post-transplantation. Clinical, biological, and histological data were collected prospectively.

The study was designed as a prospective, single-center, observational, and exploratory study. Data not normally distributed were presented as median with interquartile range (IQR: 25%–75%). Univariate analyses were performed: chi-square tests were used for categorical variables, and Kruskal–Wallis tests were used for continuous variables. T-tests or Mann–Whitney tests were used for multiple comparisons across time points. To assess whether each biomarker was associated with postoperative graft dysfunction, receiver operating characteristic (ROC) curve analyses were performed, including the calculation of positive predictive value (PPV), negative predictive value (NPV), sensitivity (Sen), and specificity (Spe). For all analyses, a p-value ≤ 0.05 was considered statistically significant; adjusted p-values are reported when data were corrected for multiple comparisons. All statistical analyses were performed using R software (version 2022, Posit Software, PBC).

Between April 2022 and April 2023, 31 patients were included. The median age of the recipients was 55 years (48–63). The MELD score was 16 (8–26) (Supplementary Table S1).

The median age of the donors was 67 years (52–74). The median BAR score was 7 (3–11), and 24 out of 31 donors (77%) met the criteria for ECDs (Supplementary Table S1).

Between liver graft procurement and reperfusion, Synd-1 levels increased significantly [9,035 (5,140–11,774) to 74,577 (55,282–107,969) pg/mL, p < 0.001]. Heparan sulfate and Ang-1 decreased significantly (heparan sulfate: 806 (453–1,056) vs. 383 (209–844) ng/mL, p < 0.01; Ang-1: 2,732 (992–6,697) vs. 1.361 (651–3,705) pg/mL, p < 0.01). Ang-2 level increased, but the change did not reach statistical significance [5,177 (3,676–8,124) vs. 6,669 (3,731–9,281) pg/mL] (Table 1).

Regarding cell injury biomarkers, levels of total CK-18 [2,689 (1,697–4,048) vs. 1,088 (402–2,188) UI/mLl, p < 0.01], cleaved CK-18 [248 (192–347) vs. 165 (98–229) UI/mL, p < 0.01], and necrosis [2,342 (1,616–3,591) vs. 1,088 (251–1,868) UI/mL, p < 0.01] were significantly higher at reperfusion compared to the procurement time point. TNF-alpha levels remained stable in the graft vena cava effluent (Figure 1).

Twelve patients (39%) developed severe graft dysfunction. Among these, 67% received a graft from a donor who died of anoxia (p = 0.001), and they exhibited significant cytolysis [ALT: 117 (22–166) vs. 27 (16–40) IU/L, p = 0.019]. No significant differences were observed between patients who developed dysfunction and those who did not regarding recipient age, MELD score, or BAR score (Table 2). Only Synd-1 levels in the graft vena cava effluent at the time of procurement were significantly higher in patients who developed dysfunction: [12,173 (10,538–17,570) vs. 6,282 (4,604–10,002) pg/mL, p = 0.004]. These patients also exhibited significantly higher necrosis levels in the graft vena cava effluent at the end of cold ischemia [1,382 (18–2,622) vs. 0 (0–0) IU/L, p = 0.004] (Table 3, Figure 2).

HS, Ang-1, and Ang-2 levels did not differ depending on the presence or absence of dysfunction.

Only a few variables differed significantly between the two groups (severe dysfunction vs. non-severe dysfunction), limiting the relevance of a multivariate analysis.

The Synd-1 threshold predicting dysfunction was 9.419.7 pg/mL [AUC = 0.81, IC 95% (0.65–0.96), sensitivity = 83%, specificity = 74%, PPV = 0.67, NPV = 0.88, p < 0.05] (Figure 3). Liver grafts with Synd-1 ≥9.419.7 pg/mL at the time of procurement exhibited significantly higher cytolysis Aspartate Aminotransférase (ASAT) 77 (56–147) vs. 26 (21–52) IU/L p < 0.001; Alanine Aminotransférase (ALAT) 58 (32–150) vs. 20 (12–40) IU/L p = 0.004] and cholestasis [GGT 52 (34–148) vs. 28 (15–52) IU/L, p = 0.013]. No significant associations were observed between glycocalyx degradation and either blood loss volume or blood product transfusion during LT (Table 4and Supplementary Table S2).

Since only Synd-1 was elevated, we assessed the occurrence of complications beyond the cut off identified on the ROC curve. Recipients who received liver graft with Synd-1 level ≥9,419.7 pg/mL in the vena cava effluent at the time of procurement experienced significantly more severe post-LT complications according to the Dindo-Clavien classification (67% vs. 19%, p = 0.007) (TABLE 4).

Despite elevated Synd-1 levels, liver grafts did not show evidence of severe histological injury after graft reperfusion (Table 4).

The sample size in our study remains small and might not be robust enough to detect correlations; therefore, the results should be interpreted with caution. In addition, for cost reasons, we only measured one inflammatory cytokine, which showed little variation, contrary to findings described in the literature. Also, sampling was limited to 1 h after portal reperfusion of the graft, preventing analysis of biomarkers in the days following transplantation. The possible impact of medications on the glycocalyx must also be considered; particularly, glucocorticoids, potent anti-inflammatory agents, have been shown to exert protective effects on the endothelium (38, 39).

Finally, we have a liver transplantation with a DCD graft, which may represent a confounding factor in the study due to differences in procurement conditions. It would be useful to perform additional analyses involving a larger number of DCD donors.

We have shown that measuring glycocalyx degradation markers as early as organ procurement can predict future liver graft function. Such early measurements could provide objective biological support to assist in graft selection. This study highlights the importance of glycocalyx preservation.