Original Clinical ScienceçLiver
Alterations of Endothelial Glycocalyx During Orthotopic Liver Transplantation in Patients With End-Stage Liver Disease Judith Schiefer, MD,1 Diana Lebherz-Eichinger, MD,1 Gabor Erdoes, MD,1 Gabriela Berlakovich, MD,2 Andreas Bacher, MD,1 Claus G. Krenn, MD,1 and Peter Faybik, MD1 Background. Endothelial glycocalyx participates in the maintenance of vascular integrity, and its perturbations cause capillary leakage, loss of vascular responsiveness, and enhanced adhesion of leukocytes and platelets. We hypothesized that marked shedding of the glycocalyx core protein, syndecan-1, occurs in end-stage liver disease (ESLD) and that it increases during orthotopic liver transplantation (OLT). We further evaluated the effects of general anesthesia on glycocalyx shedding and its association with acute kidney injury (AKI) after OLT. Patients and Methods. Thirty consecutive liver transplant recipients were enrolled in this prospective study. Ten healthy volunteers served as a control. Acute kidney injury was defined by Acute Kidney Injury Network criteria. Results. Plasma syndecan-1 was significantly higher in ESLD patients than in healthy volunteers (74.3 ± 59.9 vs 10.7 ± 9.4 ng/mL), and it further increased significantly after reperfusion (74.3 ± 59.9 vs 312.6 ± 114.8 ng/mL). The type of general anesthesia had no significant effect on syndecan-1. Syndecan-1 was significantly higher during the entire study in patients with posttransplant AKI stage 2 or 3 compared to patients with AKI stage 0 or 1. The area under the curve of the receiver operating characteristics curve of syndecane-1 to predict AKI stage 2 or 3 within 48 hours after reperfusion was 0.76 (95% confidence interval, 0.57–0.89, P = 0.005). Conclusions. Patients with ESLD suffer from glycocalyx alterations, and ischemia-reperfusion injury during OLT further exacerbates its damage. Despite a higher incidence of AKI in patients with elevated syndecan-1, it is not helpful to predict de novo AKI. Volatile anesthetics did not attenuate glycocalyx shedding in human OLT.
(Transplantation 2015;99: 2118–2123)
V
ascular endothelium is critically involved in many steps of tissue damage originating from cold and warm noflow ischemia during organ procurement, preservation, and implantation followed by re-establishment of blood flow during reperfusion in human orthotopic liver transplantation (OLT).1,2 The healthy vascular endothelium is coated with a glycocalyx, a tight and negatively charged meshwork, extending in some vessels up to 2 μm from the endothelial cell membrane.3 It consists of core proteoglycans carrying highly sulphated glycosaminoglycan attachments, and together Received 15 July 2014. Revision requested 23 November 2014. Accepted 20 December 2010. 1
Department of Anesthesiology, General Intensive Care and Pain Medicine, Medical University of Vienna, Vienna, Austria.
2
Division of Transplantation, Department of Surgery, Medical University of Vienna, Vienna, Austria. Financial support for this work was provided by departmental funds. Authors declare no conflicts of interest. J.S. and P.F. participated in research design, performance of the research, data analysis and writing of the paper. D.L.-E. and G.E. participated in performance of the research and writing of the paper. G.B., A.B., and C.G.K. participated in writing of the paper. Correspondence: Peter Faybik, MD, EDIC, Department of Anesthesiology, General Intensive Care and Pain Medicine Medical University of Vienna, Waehringer Guertel 18–20 1090 Vienna, Austria. ([email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0041-1337/15/9910-2118 DOI: 10.1097/TP.0000000000000680
2118
www.transplantjournal.com
with plasma proteins, it forms a so-called endothelial surface layer (ESL).3,4 There is increasing scientific evidence that an intact glycocalyx maintains vascular permeability, mediates stress-dependent nitric oxide production, modulates inflammatory response, and houses vascular protective enzymes as well as coagulation inhibition factors.5 In clinical settings of liver transplantation, ischemia-reperfusion injury (IRI) leads to the activation of liver Kupffer cells, production of cytokines and chemokines, generation of reactive oxygen species, increased expression of adhesion molecules, and infiltration by circulating lymphocytes and monocytes.2 Previous animal and human studies showed that IRI causes massive degradation or even destruction of glycocalyx skeleton, measured as plasma increase of its core proteins, mainly syndecan-1 and heparan sulphate.6-8 This leads to capillary leakage, edema formation, loss of endothelium-dependent vascular responsiveness, and enhanced adhesion of leukocytes and platelets, as well as the activation of coagulation pathways. Because the degree of glycocalyx destruction was positively related to organ dysfunction and overall mortality in different clinical settings, its protection and/or regeneration appears to be a promising therapeutic goal.9,10 Endothelial protection by preconditioning and/or postconditioning during IRI using different pharmacological agents was already investigated and found beneficial under controlled experimental settings.3 Both hydrocortisone and antithrombin alleviated glycocalyx shedding and polymorphonuclear neutrophils adhesion after IRI in isolated guinea pig hearts.11,12 Furthermore, hydrocortisone reduced postischemic oxidative stress, Transplantation
■
October 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
■
Volume 99
■
Number 10
Schiefer et al
© 2015 Wolters Kluwer
coronary perfusion pressure, and transudate formation after 20 minutes of warm IRI. Electron microscopy revealed a mostly intact glycocalyx skeleton after hydrocortisone treatment.12 Volatile anesthetics, especially preconditioning and/ or rapid postconditioning with sevoflurane, protected glycocalyx from IRI-induced shedding and reduced cell adhesion in animal models.6,13 Another study in 5 healthy volunteers demonstrated that sevoflurane inhalation at even subanesthetic concentrations preserves postocclusive blood flow, representing in vivo human evidence of endothelial protection against IRI by volatile anesthetics.14 Alterations in the composition and thickness of glycocalyx after exposure to an inflammatory insult, quantified by the measurement of concentrations of plasma's main components and by electron microscopic imaging, are not only involved in the pathophysiology of IRI, but also in vascular dysfunction of infection, after major surgery, diabetes, and arteriosclerosis. Activation of endothelium with compositional and dimensional changes of glycocalyx in patients with end-stage renal disease and its recovery after successful kidney transplantation was described recently.15 However, data on possible alterations of glycocalyx in patients with endstage liver disease (ESLD) are still lacking. Pathophysiological changes in arterial and venous resistance, increased permeability, oedema formation, shunting and endothelial dysfunction of splanchnicus, renal, pulmonary and cerebral microcirculation are hallmarks of ESLD and are clinically mirrored by the development of complications of liver cirrhosis. The aim of this study was to evaluate glycocalyx alterations in patients with ESLD before and during OLT, compared to healthy volunteers. We hypothesized that marked shedding of glycocalyx occurs in patients with ESLD and that it further increases due to IRI during OLT. We further evaluated the effects of general anesthesia with volatile anesthetics versus propofol-based total intravenous anesthesia on glycocalyx shedding, as well as the value of glycocalyx core protein syndecan-1 as an alternative marker of acute kidney injury (AKI) after OLT. PATIENTS AND METHODS After approval by the Ethics Committee of the Medical University of Vienna (EK Protocol Nr.: 873/2011) and given written informed consent, 30 consecutive liver transplant recipients suffering from ESLD were enrolled in the study. The study was designed as a prospective single-center observational study. Orthotopic liver transplantation was performed following the local standard technique with cross-clamping of the inferior vena cava without the use of venovenous-bypass. All grafts were ABO identical and matched for the physiological body-to-weight ratio of recipients. Immunosuppression was started intraoperatively with 40 mg of dexamethasone. The postoperative immunosuppressive protocol consisted of dexamethasone (32 mg, day 1, tapered 8 mg/day), antithymocyteglobulin (1 mg/kg per day, days 1–3), and cyclosporine (5 mg/kg per day divided in 2 doses, day 3). Data collection included the following: demographic data, underlying condition leading to OLT, surgical report, use of blood products, model of end-stage liver disease (MELD) score, laboratory parameters (blood chemistry, coagulation, full blood count) before and on the first 7 days after OLT,
2119
the need for renal replacement therapy. Preoperative creatinine clearance was estimated using the method described by the Modification of Diet in Renal Disease Study Group.16 The shedding of glycocalyx was assessed by the measurement of plasma syndecan-1 concentration at 5 different time points (TP) during and after OLT. Arterial blood samples were drawn from a radial artery catheter. Measurement of basal plasma concentration of syndecan-1 was obtained under stable anesthetic conditions after the induction of anesthesia before surgical incision (TP 0), after clamping of the inferior cava (TP 1), 10 minutes after reperfusion of the liver graft (TP 2), at the end of surgery (TP 3), and 24 hours after reperfusion (TP 4). Immediately after collection, plasma ethylenediaminetetraacetic acid samples were spun at 2600g for 15 minutes to remove particulates. Because the samples were not analyzed shortly after collection, they were stored at −80° C until analysis. Syndecan-1 concentration was determined in patient's plasma by enzyme-linked immunoassay, according to the manufacturer's information (Diaclone Research; Besancon, France). Blood samples of 10 healthy volunteers with no medical history of liver or kidney disease served as a control group. The incidence of postoperative AKI was assessed according to the Acute Kidney Injury Network (AKIN) criteria, recognizing that even small changes in serum creatinine (≥0.3 mg/dL) within 48 hours adversely impact clinical outcome.16,17 The AKIN definition incorporates serum creatinine, urine output, and time. According to these criteria, there are 3 stages of AKI. Stage 1: an increase in serum creatinine of 0.3 mg/dL or higher or a 1.5- to 2-fold increase from baseline and/or urine output less than 0.5 mL/kg per hour for longer than 8 hours; stage 2: a 2- to 3-fold increase in serum creatinine from baseline and/or urine output less than 0.5 mL/kg per hour for longer than 12 hours; stage 3: a 3-fold increase in serum creatinine from baseline or serum creatinine of 4 mg/dL or higher with an acute rise of at least 0.5 mg/dL and/or urine output less than 0.5 mL/kg per hour for 24 hours or anuria for 12 hours. Data are presented as mean ± standard deviation, unless indicated otherwise. The Gaussian distribution of samples was tested with the D'Agostino-Pearson test. One-way analysis of variance was used to test the difference between the means of several subgroups of a variable (multiple testing) and post hoc Student-Newman-Keuls test was used for pairwise comparison of subgroups, when analysis of variance test was positive. Correlation analysis was performed according to Pearson. A receiver operating characteristics (ROC) curve analysis was used for syndecan-1 as a marker of AKI to plot the true-positive rate in function of the false-positive rate at different cutoff points and area under the curve (AUC), sensitivity, and specificity were calculated. Sample size calculation for area under ROC curve of 0.8 with type I error of 0.05 and type II error of 0.2 (power 80%) estimated a minimal sample size of 26 patients. All statistical analyses were performed with the MedCalc Statistical Software version 13.0.4 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org, 2014). RESULTS Thirty consecutive patients (21 were men) undergoing OLT were consecutively enrolled in the study. The demographic
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2120
Transplantation
■
October 2015
■
Volume 99
■
Number 10
data and type of ESLD leading to OLT are summarized in Table 1. Baseline preoperative laboratory parameters and postoperative laboratory parameters after OLT are shown in Table 2. Basal plasma syndecan-1 concentration obtained under stable anesthetic conditions after induction of anesthesia before surgical incision (TP 0) was significantly higher in liver transplant recipients than in healthy volunteers (Table 3). Furthermore, plasma levels of syndecan-1 changed significantly during and after OLT (analysis of variance, P < 0.001). There was a significant increase in plasma levels of syndecan-1 after reperfusion (TP 2) compared to basal plasma concentration (TP 0) and clamping of the inferior vena cava (TP 1) (post hoc Student-Newman-Keuls test, P < 0.05). This significant increase persisted at the end of surgery (TP 3) and 24 hours after reperfusion (TP 4) (Figure 1). We found a significant correlation between increased plasma syndecan-1 levels after reperfusion (TP 2) and aspartate aminotransferase levels on day 1 (P = 0.01; Pearson correlation coefficient r = 0.45) and day 2 (P = 0.02; Pearson correlation coefficient r = 0.4) after OLT but not on day 3 (P = 0.3; Pearson correlation coefficient r = 0.3). Mean cross-clamping time of the inferior vena cava (warm ischemia time) took 76 minutes (range, 50–100 minutes); the average duration of cold ischemia time was 425.7 minutes (range, 200–880 minutes). There was no significant correlation between warm (P = 0.4; Pearson correlation coefficient r = 0.15) or cold ischemia time (P = 0.6; Pearson correlation coefficient r = 0.1) and an increase in plasma syndecan-1 levels during and after OLT. The mean ICU and hospital length of stay were 15 ± 16 days (range, 2–70 days) and 30 ± 29 days (range, 9–148 days), respectively. We found a significant correlation between ICU and hospital length of stay (P < 0. 0001; Pearson correlation coefficient 0.84). Twenty-eight patients (93.4%) survived 28 days after OLT. One patient died due to hemorrhage and consecutive multiorgan failure, the other due to sepsis. Ninety-day and 1-year survival rates were 90% and 84%, respectively. One patient died in septic shock after
TABLE 1.
Demographic data and indications for liver transplantation Demographic data
Age, y Sex (male/female) Body surface area, m2 Child-Pugh score (A/B/C) MELD score Etiology of liver disease: Alcohol Viral Alcohol and hepatocellular carcinoma Viral and hepatocellular carcinoma Cholestatic liver disease Wilson disease Cryptogenic cirrhosis Autoimmune cirrhosis Other
54 ± 11 21 (70 %) / 9 (30 %) 1.9 ± 0.2 1 / 13 / 16 16 ± 6.9 6 (20 %) 8 (26 %) 1 (3 %) 6 (20 %) 2 (7 %) 3 (10 %) 1 (3 %) 1 (3 %) 2 (7 %)
Results are presented as mean ± standard deviation or number and percentage.
www.transplantjournal.com
re-OLT because of hepatic artery thrombosis on day 42, the other 2 patients secondary to sepsis. Plasma syndecan-1 levels after reperfusion (TP 2) were significantly lower in patients who survived 28 days (survivors 300 ± 108 ng/mL vs nonsurvivors 483 ± 50 ng/mL; P = 0.02) and 90 days (survivors 296 ± 108 ng/mL vs nonsurvivors 457 ± 56 ng/mL; P = 0.01) after OLT than in those who died. There was no significant difference in plasma syndecan-1 levels at TP 2 between patients who survived 1 year after OLT and those who died (P = 0.06). Pretransplant syndecan-1 levels were not significantly different in respect of survival after OLT. Fifteen liver graft recipients developed AKI stage 2 (n = 2) or 3 (n = 12) based on the AKIN criteria, and 10 of them were supported with continuous renal replacement therapy (CRRT). The mean duration of CRRT was 5 ± 10 days (range, 0–40 days). Furthermore, 3 of these patients underwent subsequent 1 to 5 treatments on intermittent hemodialysis before their renal function recovered. Three graft recipients developed AKI stage 1, and 12 graft recipients had no AKI, none of these required CRRT. Patients with AKI stage 2 or 3 had significantly higher plasma levels of syndecan-1 during the entire study period, compared to patients with AKI stage 0 or 1 (Table 3). We found a significant negative correlation between plasma syndecan-1 levels before surgery (TP 0) and preoperative estimated glomerular filtration rate (78.7 ± 29.3 mL/minute) with the Modification of Diet in Renal Disease formula (P = 0.002; Pearson correlation coefficient r = − 0.53). Although not significant, the mean pretransplant serum creatinine in patients who developed AKI stage 0 or 1 was lower than in patients who developed AKI 2 or 3 (0.96 ± 0.38 mg/dL vs 1.27 ± 0.49 mg/dL, respectively, P = 0.06). However, there was no significant difference in the increase of plasma levels of syndecan-1 from TP 1 to TP 2 between the patients with AKI stage 0 or 1 and patients with AKI stage 2 or 3 (unpaired t test, P = 0.19). Similarly, the mean pretransplant model of endstage liver disease score in patients who developed AKI stage 0 or 1 was lower than that in patients who developed AKI 2 or 3 (13.6 ± 6.2 vs 18.3 ± 7.0, respectively, P = 0.06). In the ROC curve analysis, syndecan-1 at TP 2 and δ syndecan-1 (TP 2 minus TP 0) exhibited an AUC of 0.76 (95% confidence interval, 0.57-0.89; P = 0.005) and 0.64 (95% confidence interval, 0.45-0.80; P = 0.17), respectively, for the development of AKI stage 2 or 3 within 48 hours after graft reperfusion. The sensitivity and specificity for syndecan-1 at TP 2 with optimal cutoff value greater than 379 ng/mL were 60% and 100%, respectively. The sensitivity and specificity for δ syndecan-1 with optimal cutoff value greater than 314 ng/mL were 40 % and 93 %, respectively. General anesthesia was maintained with volatile anesthetics (sevoflurane, n = 7; desflurane, n = 5) or with propofol-based total intravenous anesthesia (n = 18). The type of general anesthesia had no significant effect on plasma levels of syndecan-1 (unpaired t test; P = 0.28). There was no difference between patients anesthetized with sevoflurane and desflurane in regard to syndecan-1 levels. DISCUSSION This study is the first to describe the perioperative changes of plasma syndecan-1 as a marker of endothelial glycocalyx damage during human OLT. Notably, the baseline plasma
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2121
Schiefer et al
© 2015 Wolters Kluwer
TABLE 2.
Baseline and postoperative laboratory parameters before and after liver transplantation in 30 patients with end-stage liver disease Parameter
Sodium, mmol/L Crea, mg/dL Bilirubin, mg/dL AST, U/L ALT, U/L Hb, mg/dL WBC, G/L Plt, G/L PT (%)
pre-OLT
POD 1
POD 2
POD 3
POD 7
136 ± 4 1.1 ± 0.4 4.7 ± 6 123 ± 157 112 ± 265 11.2 ± 1.6 5.7 ± 2.5 98.6 ± 54.5 56 ± 20
138 ± 4 1.26 ± 0.4 5.0 ± 3.1 1589 ± 1198a 849 ± 526 9.5 ± 1.2 11.4 ± 4.4 90.0 ± 74.4 46 ± 13
139 ± 4 1.63 ± 0.7 5.9 ± 4.3 1329 ± 1506* 776 ± 581 9.2 ± 1.4 10.9 ± 4.9 63.0 ± 64.1 43 ± 12
138 ± 4 1.69 ± 0.9 5.1 ± 3.6 644 ± 723 648 ± 548 8.8 ± 1.1 6.9 ± 3.9 42.3 ± 30.9 51 ± 14
137 ± 4 1.26 ± 0.5 5.3 ± 5.0 278 ± 1057 304 ± 545 9.3 ± 1.2 8.4 ± 6.1 61.1 ± 37.2 69 ± 12
a
A significant correlation between AST and plasma syndecan-1 levels after reperfusion due to IRI (P < 0.05). Results are presented as mean ± standard deviation. POD, postoperative day; Crea, creatinine; AST, aspartate aminotransferase; ALT, alanine aminotransferase; Hb, hemoglobin; WBC, white blood cells; Plt, platelet count; PT, prothrombin time.
syndecan-1 before surgery was significantly higher (7-fold) in liver transplant recipients compared to healthy volunteers, reflecting the pathophysiological alteration of glycocalyx in ESLD patients. Although no data on syndecan-1 in ESLD patients were previously published, the baseline syndecan-1 levels in our transplant recipients before OLT were comparable to the plasma syndecan-1 concentrations recently described after major abdominal surgery.9 The pronounced shedding of syndecan-1 before OLT may result from the derangement of inflammatory response seen in patients with liver cirrhosis with consequent overexpression of various proinflammatory and anti-inflammatory cytokines.18 Furthermore, perioperative proinflammatory cytokine levels and permeability of sinusoids in cirrhotic livers were shown to correlate with the severity of the liver disease necessitating OLT.11,18,19 Among other possible pathophysiological mechanisms, bacterial translocation, often occurring in ESLD, is an important trigger of systemic inflammation and may thus potentiate vascular dysfunction in ESLD.20 Bacterial lipopolysaccharide and tumor necrosis factor-α were both shown to cause glycocalyx shedding.21,22 In line with our results in ESLD patients, compositional and dimensional changes of glycocalyx in sublingual microvasculature were recently described in patients with end-stage renal disease.15 Reduced renal function was strongly associated with low ESL dimension. Although not yet reported in human OLT, the significant increase of plasma syndecan-1 after reperfusion in our patients is in accordance with data from animal and human
studies investigating glycocalyx shedding after IRI. Major degradation of glycocalyx was described after 20 minutes in an isolated heart model and after 60 minutes of warm ischemia followed by reperfusion in a less sensitive bowel model.7 Human studies confirmed these results and showed a concomitant increase of glycocalyx degradation products with reperfusion.8,23 Notably, plasma syndecan-1 after reperfusion in our collective reached concentrations as high as previously described in septic patients,9 and was 29.4-fold higher than in healthy controls. Furthermore, levels of proinflammatory markers, such as interleukin-6, were correlated with syndecan-1 levels in septic patients.9 This is in line with our previous results in patients with ESLD undergoing OLT without venovenous bypass showing significant increase of interleukin-6 after reperfusion.24 There was also a significant positive correlation between syndecan-1 after reperfusion and serum aspartate aminotransferase on the first 2 postoperative days, reflecting the damage caused by IRI. Although transaminase activity levels represent a well-established parameter for the estimation of hepatic injury after OLT, their increased levels alone are not predictive of outcome or incidence of early allograft dysfunction. Because of a small sample size, we cannot conclude about the predictive value of syndecan-1 alone on early mortality after OLT, but plasma syndecan-1 levels after reperfusion were significantly lower in patients who survived 28 and 90 days after OLT than in those who died. Because pretransplant syndecan-1 levels were not significantly different with respect of survival after OLT, factors influencing integrity of glycocalyx, such as
TABLE 3.
Plasma levels of syndecan-1 (ng/mL) in human liver transplantation
Control TP 0 TP 1 TP 2 TP 3 TP 4 δ syndecan-1
All patients n = 30
AKIN 0 – 1 n = 15
AKIN 2 – 3 n = 15
p
10.6 ± 9.4 74.3 ± 59.9 83.5 ± 53.3 312.6 ± 114.8 291.8 ± 117.5 279.7 ± 138.3 238.3 ± 122.5
49.3 ± 50.1 60.4 ± 52.9 257.7 ± 82.1 248.1 ± 97.6 202.8 ± 88.1 208.4 ± 86.8
99.1 ± 59.8 106.7 ± 44.1 367.4 ± 118.9 335.4 ± 122.4 351.5 ± 140.0 268.2 ± 146.9
0.018 0.015 0.006 0.039 0.002 0.19
δ syndecan-1 represents difference between TP 0 and TP 2, Results are presented as mean ± standard deviation. P value applies for comparison of plasma syndecan-1 concentration in patients with AKI stage 0 or 1 versus AKI stage 2 and 3.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2122
Transplantation
■
October 2015
■
Volume 99
■
Number 10
Figure 1. Measurements of circulatory endothelial surface layer damage marker syndecan-1 in patients with end-stage liver disease undergoing liver transplantation. *TP 0 and TP 1 versus healthy volunteers (P < 0.05); **TP 2, TP 3, and TP 4 versus TP 0 (P < 0.05).
IRI, quality of graft and ischemia time, could have potentially influenced the outcome of our patients. Acute kidney injury after OLT is a recognized independent risk factor for morbidity, mortality, reduced graft survival, longer hospital stay, and higher costs.17,25 Recent studies using a consensus definition based on the AKIN showed that when AKI is defined as at least a doubling of serum creatinine levels or the need for RRT (AKIN stages 2 or 3), the incidence of AKI in OLT is up to approximately 48%.17 This is in agreement with our results because we found AKI stage 2 or 3 in 50% of our patients and 33% of these received CRRT. However, serum creatinine is slow and insensitive because it does not immediately reflect the dynamic changes in renal function and therefore does not alert the clinician of impeding renal injury. In our study, AUC of the ROC analysis of plasma syndecan-1 and increase of syndecane-1 after reperfusion to predict AKI (stage 2 or stage 3) was 0.75 and 0.64, respectively. However, because of the low sensitivity, plasma syndecan-1 is not helpful as an alternative marker of de novo AKI in human OLT. In a study from Wagener et al,25 urinary neutrophil gelatinase-associated lipocalin-tourine creatinine ratio demonstrated encouraging results with much better discrimination power to predict AKI stage 3 after OLT. According to our results, patients who developed AKI stage 2 or 3 had significantly higher plasma syndecan-1 compared to graft recipients with AKI stage 0 or 1 before OLT, and it remained significantly higher throughout the whole study. Furthermore, a significant negative correlation between estimated glomerular filtration rate and syndecan-1 before surgery was found in our patients. Serum creatinine was also slightly higher before OLT (with a mean difference of 0.3 mg/dL) in patients who developed AKI stage 2 or 3 compared to those with AKI stage 0 or 1 after OLT. Bearing in mind that all transplant recipients in this study had, compared to healthy volunteers, significantly higher plasma levels of syndecan-1 before OLT, these results suggest that also slightly elevated or nearly normal serum creatinine (due to malnutrition and reduced muscle mass) in combination with pre-OLT increased syndecan-1, reflects a higher grade of chronic inflammatory injury of ESLD predisposing to AKI and higher mortality. Reduced renal clearance of syndecan-1 could also be involved in its increased plasma
www.transplantjournal.com
concentration, but its large molecular weight excludes a direct renal excretion. Although exact clearance of syndecan-1 is still unknown, low levels of syndecan-1 were shown in patients with interstitial fibrosis/tubular atrophy after kidney transplantation and reduced renal function.15 Our results are in line with a previous study by Dane et al15 that showed a strong association of reduced renal function with low endothelial surface dimensions in patients with end-stage renal disease. After successful kidney transplantation, the ESL of the sublingual microvasculature, measured by a novel noninvasive sidestream-darkfield imaging model, was indistinguishable from healthy control. Similarly, successful OLT often leads to improvement of reduced renal function seen in ESLD or even reversal of hepatorenal syndrome. Although 10 patients in our study were supported with CRRT and 3 patients later with intermittent hemodialysis, none of them required extracorporeal renal support at the end of follow-up. The negative impact of IRI on the integrity of glycocalyx contributing to endothelial dysfunction makes its protection and/or restoration a valuable therapeutic goal, and preconditioning and postconditioning with various pharmacological agents in setting of IRI was recently studied.3,6,14,26 Lucchinetti et al14 reported a markedly improved postocclusive hyperaemic reaction and attenuated activation of leukocytes with peri-ischemic inhalation of sevoflurane, targeted at 0.5 to 1 vol% end-tidal concentrations, after 15 minutes of forearm ischemia followed by reperfusion in 5 healthy volunteers. The targeted end-tidal concentration of sevoflurane in this study was rather low, but approximates the end-tidal concentration of sevoflurane used in our study during cross-clamping of the inferior vena cava and thus warm ischemia and early reperfusion phase of human OLT. Furthermore, in an isolated guinea pig heart model, inhalation of 1 minimal alveolar concentration sevoflurane protected the endothelial glycocalyx from IRI-induced degradation, with both preconditioning and rapid postconditioning being successful.6 The postconditioning (during reperfusion period) in this study mimics the clinical setting of human OLT because in case of volatile anesthetics, the liver graft is exposed to sevoflurane first during reperfusion. Contrary to these experimental studies, we observed no differences in plasma syndecan-1 at any time during the OLT, regardless of the use of volatile anesthetics or propofol-based anesthesia. This may be due to cold ischemia time that was not part of the experimental studies but is currently unavoidable in real-life clinical transplantation and considerably longer warm ischemia than in the studies mentioned above. Because we did not perform a power analysis on the effect of volatile anesthetics on glycocalyx shedding, we cannot exclude any beneficial effects of volatile anesthetics in a larger study population, but the previous studies found significant differences in similar or even smaller-sized study populations. In line with the lack of protective effects of volatile anesthetics on glycocalyx shedding in our study, the beneficial effects of corticosteroids, assessed previously by reduced transudate formation, leukocyte adhesion, and decreased shedding of syndecan-1 after IRI in animal models, were again in contrast with our results.12,27 According to our immunosuppressive protocol, 40 mg of dexamethasone was intravenously injected before cross clamping of the inferior vena cava. There was no steroid in the preservation or flushing solution because it was part of the experimental protocol
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Schiefer et al
© 2015 Wolters Kluwer
in previously mentioned successful animal models.12,27 Our data are in line with the results reported by Biancofiore et al,18 where the administration of cortisone during OLT as a part of immunosuppressive therapy did not change the inflammatory pattern nor affected the cytokine balance by shifting it toward inflammation or over-immunosuppression. Based on these major discrepancies, no conclusion can be drawn on the potentially beneficial effect of steroids on glycocalyx shedding in clinical settings without adding steroids to preservation solutions. However not unexpected, 40 mg of dexamethasone intravenously before cross-clamping of the caval vein in human OLT definitively does not prevent glycocalyx from pronounced shedding. Similar to this model, augmenting histidine-tryptophan-ketoglutarate solutions with human albumin improved endothelial integrity and heart performance after 4 hours of cold ischemia of guinea pig hearts.28 Although we routinely flushed all liver grafts with 20% human albumin before reperfusion for at least 20 minutes, massive syndecan-1 increase has not been attenuated. The different timing of albumin administration most probably hinders the translation of beneficial effects seen under experimental conditions into clinical setting and makes further clinical investigations necessary. Our study has several limitations. Because we collected our blood samples from the radial artery, only systemic changes in syndecan-1 could be investigated. The flushing of liver grafts with human albumin before reperfusion could wash out some syndecan-1 molecules. The contribution of OLT as major abdominal surgery additionally to IRI-induced glycocalyx shedding cannot be distinguished under these circumstances. Furthermore, cross-clamping of the inferior vena cava with consecutive renal and splanchnic venous congestion in OLT, without the use of venovenous bypass, is at least in part responsible for the pronounced syndecan-1 increase seen after reperfusion. In conclusion, plasma syndecan-1 concentration as a marker of glycocalyx shedding in patients with ESLD is significantly increased when compared to healthy population. Ischemia-reperfusion injury during OLT further exacerbates glycocalyx shedding. Despite a higher incidence of AKI and the need for CRRT in patients with higher syndecan-1 concentration, syndecan-1 is not helpful as an alternative biomarker to predict de novo AKI. Contrary to previous experimental and animal studies, rapid postconditioning with volatile anesthetics did not attenuate glycocalyx shedding under clinical condition of human OLT. REFERENCES 1. Seal JB, Gewertz BL. Vascular dysfunction in ischemia-reperfusion injury. Ann Vasc Surg. 2005;19:572–584. 2. Zhai Y, Petrowsky H, Hong JC, Busuttil RW, Kupiec-Weglinski JW. Ischaemia-reperfusion injury in liver transplantation—from bench to bedside. Nat Rev Gastroenterol Hepatol. 2013;10:79–89. 3. Becker BF, Chappell D, Bruegger D, Annecke T, Jacob M. Therapeutic strategies targeting the endothelial glycocalyx: acute deficits, but great potential. Cardiovasc Res. 2010;87:300–310. 4. Salmon AHJ, Satchell S. Endothelial glycocalyx dysfunction in disease: albuminuria and increased microvascular permeability. J Pathol. 2012; 226:562–574. 5. Weinbaum S, Tarbell JM, Damiano ER. The structure and function of the endothelial glycocalyx layer. Annu Rev Biomed Eng. 2007;9:121–167.
2123
6. Annecke T, Chappell D, Chen C, et al. Sevoflurane preserves the endothelial glycocalyx against ischaemia-reperfusion injury. Br J Anaesth. 2010;104:414–421. 7. Mulivor AW, Lipowsky HH. Inflammation- and ischemia-induced shedding of vebular glycocalyx. Am J Physiol Heart Circ Physiol. 2004;286: H1672–H1680. 8. Rehm M, Bruegger D, Christ F, Conzen P, Thiel M, Jacob M. Shedding of endothelial glycocalyx in patients undergoing major vascular surgery with global and regional ischemia. Circulation. 2007;116:1896–1906. 9. Steppan J, Hofer S, Funke B, et al. Sepsis and major abdominal surgery lead to flaking of the endothelial glycocalix. J Surg Res. 2011; 165:136–141. 10. Nelson A, Beckerstedt I, Schmidtchen A, Ljunggren L, Bodelsson M. Increased levels of glycosaminoglycans during septic shock: relation to mortality and the antimicrobial actions of plasma. Shock. 2008; 30:623–627. 11. Koh HJ, Ryu KH, Cho ML, Heo YJ, Lee J. Factors influencing the concentration of cytokines during liver transplantation. Transplant Proc. 2010;42:3617–3619. 12. Chappell D, Dörfler N, Jacob M, et al. Glycocalyx protection reduces leucocyte adhesion after ischemia/reperfusion. Shock. 2010; 34: 133–139. 13. Chappell D, Heindl B, Jacob M, et al. Sevoflurane reduces leukocyte and platelet adhesion after ischemia-reperfusion by protecting the endothelial glycocalyx. Anesthesiology. 2011;115:483–491. 14. Lucchinetti E, Ambrosio S, Aguirre J, et al. Sevoflurane inhalation at sedative concentrations provides endothelial protection against ischemia-reperfusion injury in humans. Anesthesiology. 2007;106: 262–268. 15. Dane MJC, Khairoun M, Lee DH, van der Berg BM, Eskens BJM, Boels MGS. Association of kidney function with changes in the endothelial surface layer. Clin J Am Soc Nephrol. 2014;4:698–704. 16. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130:461–470. 17. Razonable RR, Findlay JY, O'Riordan A, et al. Critical care issues in patients after liver transplantation. Liver Transpl. 2011;17:511–527. 18. Biancofiore G, Bindi L, Miccoli M, et al. Balance of pro- and anti-inflammatory cytokines in cirrhotic patients undergoing liver transplantation. Transplant Immunol. 2013;28:193–197. 19. Villeneuve JP, Daganais M, Hiet PM, Lapointe R, Marleau D. The hepatic microcirculation in the isolated perfused human liver. Hepatology. 1996;23:24–31. 20. Metha G, Gustot T, Mookerjee RP, et al. Inflammation and portal hypertension—the undiscovered country. J Hepatol. 2014 http://dx.doi. org/10.1010/j.jhep.2014.03.014. 21. Henry CB, Duling BR. TNF-alpha increases entry of macromolecules into luminal endothelial cell glycocalyx. Am J Physiol Heart Circ Physiol. 2000;279:H82815–H82823. 22. Nieuwdrop M, Meuwese MC, Mooij HL, et al. Tumor necrosis factor-alpha inhibition protects against endotoxin-induced endothelial glycocalyx perturbation. Atherosclerosis. 2009;202:296–303. 23. Bruegger D, Rehm M, Abicht J, et al. Shedding of the endothelial glycocalyx during cardiac surgery: on-pump versus off-pump coronary artery bypass graft surgery. J Thorac Cardiovasc Surg. 2009; 138:1445–1447. 24. Faybik P, Hetz H, Krenn CG, Baker A, Berlakovich GA, Steltzer H. Perioperative cytokines during orthotopic liver transplantation without venovenous bypass. Transplant Proc. 2003;35:3019–3021. 25. Wagener G, Minhaz M, Mattis FA, Kim M, Emond JC, Lee HT. Urinary neuthrophil gelatinase-associated lipocalin as a marker of acute kidney injury after orthotopic liver transplantation. Nephrol Dial Transplant. 2011;26:1717–1723. 26. Smit KF, Oei GT, Brevoord D, Stroes ES, Nieuwland R, Schlack WS. Helium induces preconditioning in human endothelium in vivo. Anesthesiology. 2013;118:95–104. 27. Chappell D, Jacob M, Hofmann-Kiefer K, Bruegger D, Rehm M, Conzen P. Hydrocortisone preserves the vascular barrier by protecting the endothelial glycocalyx. Anesthesiology. 2007;107:776–784. 28. Jacob M, Mehringer PO, Chappell D, et al. Albumin augmentation improves condition of guinea pig hearts after 4 hr of cold ischemia. Transplantation. 2009;87:956–965.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Alterations of Endothelial Glycocalyx During Orthotopic Liver Transplantation in Patients With End-Stage Liver Disease Judith Schiefer, MD,1 Diana Lebherz-Eichinger, MD,1 Gabor Erdoes, MD,1 Gabriela Berlakovich, MD,2 Andreas Bacher, MD,1 Claus G. Krenn, MD,1 and Peter Faybik, MD1 Background. Endothelial glycocalyx participates in the maintenance of vascular integrity, and its perturbations cause capillary leakage, loss of vascular responsiveness, and enhanced adhesion of leukocytes and platelets. We hypothesized that marked shedding of the glycocalyx core protein, syndecan-1, occurs in end-stage liver disease (ESLD) and that it increases during orthotopic liver transplantation (OLT). We further evaluated the effects of general anesthesia on glycocalyx shedding and its association with acute kidney injury (AKI) after OLT. Patients and Methods. Thirty consecutive liver transplant recipients were enrolled in this prospective study. Ten healthy volunteers served as a control. Acute kidney injury was defined by Acute Kidney Injury Network criteria. Results. Plasma syndecan-1 was significantly higher in ESLD patients than in healthy volunteers (74.3 ± 59.9 vs 10.7 ± 9.4 ng/mL), and it further increased significantly after reperfusion (74.3 ± 59.9 vs 312.6 ± 114.8 ng/mL). The type of general anesthesia had no significant effect on syndecan-1. Syndecan-1 was significantly higher during the entire study in patients with posttransplant AKI stage 2 or 3 compared to patients with AKI stage 0 or 1. The area under the curve of the receiver operating characteristics curve of syndecane-1 to predict AKI stage 2 or 3 within 48 hours after reperfusion was 0.76 (95% confidence interval, 0.57–0.89, P = 0.005). Conclusions. Patients with ESLD suffer from glycocalyx alterations, and ischemia-reperfusion injury during OLT further exacerbates its damage. Despite a higher incidence of AKI in patients with elevated syndecan-1, it is not helpful to predict de novo AKI. Volatile anesthetics did not attenuate glycocalyx shedding in human OLT.
(Transplantation 2015;99: 2118–2123)
V
ascular endothelium is critically involved in many steps of tissue damage originating from cold and warm noflow ischemia during organ procurement, preservation, and implantation followed by re-establishment of blood flow during reperfusion in human orthotopic liver transplantation (OLT).1,2 The healthy vascular endothelium is coated with a glycocalyx, a tight and negatively charged meshwork, extending in some vessels up to 2 μm from the endothelial cell membrane.3 It consists of core proteoglycans carrying highly sulphated glycosaminoglycan attachments, and together Received 15 July 2014. Revision requested 23 November 2014. Accepted 20 December 2010. 1
Department of Anesthesiology, General Intensive Care and Pain Medicine, Medical University of Vienna, Vienna, Austria.
2
Division of Transplantation, Department of Surgery, Medical University of Vienna, Vienna, Austria. Financial support for this work was provided by departmental funds. Authors declare no conflicts of interest. J.S. and P.F. participated in research design, performance of the research, data analysis and writing of the paper. D.L.-E. and G.E. participated in performance of the research and writing of the paper. G.B., A.B., and C.G.K. participated in writing of the paper. Correspondence: Peter Faybik, MD, EDIC, Department of Anesthesiology, General Intensive Care and Pain Medicine Medical University of Vienna, Waehringer Guertel 18–20 1090 Vienna, Austria. ([email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0041-1337/15/9910-2118 DOI: 10.1097/TP.0000000000000680
2118
www.transplantjournal.com
with plasma proteins, it forms a so-called endothelial surface layer (ESL).3,4 There is increasing scientific evidence that an intact glycocalyx maintains vascular permeability, mediates stress-dependent nitric oxide production, modulates inflammatory response, and houses vascular protective enzymes as well as coagulation inhibition factors.5 In clinical settings of liver transplantation, ischemia-reperfusion injury (IRI) leads to the activation of liver Kupffer cells, production of cytokines and chemokines, generation of reactive oxygen species, increased expression of adhesion molecules, and infiltration by circulating lymphocytes and monocytes.2 Previous animal and human studies showed that IRI causes massive degradation or even destruction of glycocalyx skeleton, measured as plasma increase of its core proteins, mainly syndecan-1 and heparan sulphate.6-8 This leads to capillary leakage, edema formation, loss of endothelium-dependent vascular responsiveness, and enhanced adhesion of leukocytes and platelets, as well as the activation of coagulation pathways. Because the degree of glycocalyx destruction was positively related to organ dysfunction and overall mortality in different clinical settings, its protection and/or regeneration appears to be a promising therapeutic goal.9,10 Endothelial protection by preconditioning and/or postconditioning during IRI using different pharmacological agents was already investigated and found beneficial under controlled experimental settings.3 Both hydrocortisone and antithrombin alleviated glycocalyx shedding and polymorphonuclear neutrophils adhesion after IRI in isolated guinea pig hearts.11,12 Furthermore, hydrocortisone reduced postischemic oxidative stress, Transplantation
■
October 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
■
Volume 99
■
Number 10
Schiefer et al
© 2015 Wolters Kluwer
coronary perfusion pressure, and transudate formation after 20 minutes of warm IRI. Electron microscopy revealed a mostly intact glycocalyx skeleton after hydrocortisone treatment.12 Volatile anesthetics, especially preconditioning and/ or rapid postconditioning with sevoflurane, protected glycocalyx from IRI-induced shedding and reduced cell adhesion in animal models.6,13 Another study in 5 healthy volunteers demonstrated that sevoflurane inhalation at even subanesthetic concentrations preserves postocclusive blood flow, representing in vivo human evidence of endothelial protection against IRI by volatile anesthetics.14 Alterations in the composition and thickness of glycocalyx after exposure to an inflammatory insult, quantified by the measurement of concentrations of plasma's main components and by electron microscopic imaging, are not only involved in the pathophysiology of IRI, but also in vascular dysfunction of infection, after major surgery, diabetes, and arteriosclerosis. Activation of endothelium with compositional and dimensional changes of glycocalyx in patients with end-stage renal disease and its recovery after successful kidney transplantation was described recently.15 However, data on possible alterations of glycocalyx in patients with endstage liver disease (ESLD) are still lacking. Pathophysiological changes in arterial and venous resistance, increased permeability, oedema formation, shunting and endothelial dysfunction of splanchnicus, renal, pulmonary and cerebral microcirculation are hallmarks of ESLD and are clinically mirrored by the development of complications of liver cirrhosis. The aim of this study was to evaluate glycocalyx alterations in patients with ESLD before and during OLT, compared to healthy volunteers. We hypothesized that marked shedding of glycocalyx occurs in patients with ESLD and that it further increases due to IRI during OLT. We further evaluated the effects of general anesthesia with volatile anesthetics versus propofol-based total intravenous anesthesia on glycocalyx shedding, as well as the value of glycocalyx core protein syndecan-1 as an alternative marker of acute kidney injury (AKI) after OLT. PATIENTS AND METHODS After approval by the Ethics Committee of the Medical University of Vienna (EK Protocol Nr.: 873/2011) and given written informed consent, 30 consecutive liver transplant recipients suffering from ESLD were enrolled in the study. The study was designed as a prospective single-center observational study. Orthotopic liver transplantation was performed following the local standard technique with cross-clamping of the inferior vena cava without the use of venovenous-bypass. All grafts were ABO identical and matched for the physiological body-to-weight ratio of recipients. Immunosuppression was started intraoperatively with 40 mg of dexamethasone. The postoperative immunosuppressive protocol consisted of dexamethasone (32 mg, day 1, tapered 8 mg/day), antithymocyteglobulin (1 mg/kg per day, days 1–3), and cyclosporine (5 mg/kg per day divided in 2 doses, day 3). Data collection included the following: demographic data, underlying condition leading to OLT, surgical report, use of blood products, model of end-stage liver disease (MELD) score, laboratory parameters (blood chemistry, coagulation, full blood count) before and on the first 7 days after OLT,
2119
the need for renal replacement therapy. Preoperative creatinine clearance was estimated using the method described by the Modification of Diet in Renal Disease Study Group.16 The shedding of glycocalyx was assessed by the measurement of plasma syndecan-1 concentration at 5 different time points (TP) during and after OLT. Arterial blood samples were drawn from a radial artery catheter. Measurement of basal plasma concentration of syndecan-1 was obtained under stable anesthetic conditions after the induction of anesthesia before surgical incision (TP 0), after clamping of the inferior cava (TP 1), 10 minutes after reperfusion of the liver graft (TP 2), at the end of surgery (TP 3), and 24 hours after reperfusion (TP 4). Immediately after collection, plasma ethylenediaminetetraacetic acid samples were spun at 2600g for 15 minutes to remove particulates. Because the samples were not analyzed shortly after collection, they were stored at −80° C until analysis. Syndecan-1 concentration was determined in patient's plasma by enzyme-linked immunoassay, according to the manufacturer's information (Diaclone Research; Besancon, France). Blood samples of 10 healthy volunteers with no medical history of liver or kidney disease served as a control group. The incidence of postoperative AKI was assessed according to the Acute Kidney Injury Network (AKIN) criteria, recognizing that even small changes in serum creatinine (≥0.3 mg/dL) within 48 hours adversely impact clinical outcome.16,17 The AKIN definition incorporates serum creatinine, urine output, and time. According to these criteria, there are 3 stages of AKI. Stage 1: an increase in serum creatinine of 0.3 mg/dL or higher or a 1.5- to 2-fold increase from baseline and/or urine output less than 0.5 mL/kg per hour for longer than 8 hours; stage 2: a 2- to 3-fold increase in serum creatinine from baseline and/or urine output less than 0.5 mL/kg per hour for longer than 12 hours; stage 3: a 3-fold increase in serum creatinine from baseline or serum creatinine of 4 mg/dL or higher with an acute rise of at least 0.5 mg/dL and/or urine output less than 0.5 mL/kg per hour for 24 hours or anuria for 12 hours. Data are presented as mean ± standard deviation, unless indicated otherwise. The Gaussian distribution of samples was tested with the D'Agostino-Pearson test. One-way analysis of variance was used to test the difference between the means of several subgroups of a variable (multiple testing) and post hoc Student-Newman-Keuls test was used for pairwise comparison of subgroups, when analysis of variance test was positive. Correlation analysis was performed according to Pearson. A receiver operating characteristics (ROC) curve analysis was used for syndecan-1 as a marker of AKI to plot the true-positive rate in function of the false-positive rate at different cutoff points and area under the curve (AUC), sensitivity, and specificity were calculated. Sample size calculation for area under ROC curve of 0.8 with type I error of 0.05 and type II error of 0.2 (power 80%) estimated a minimal sample size of 26 patients. All statistical analyses were performed with the MedCalc Statistical Software version 13.0.4 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org, 2014). RESULTS Thirty consecutive patients (21 were men) undergoing OLT were consecutively enrolled in the study. The demographic
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2120
Transplantation
■
October 2015
■
Volume 99
■
Number 10
data and type of ESLD leading to OLT are summarized in Table 1. Baseline preoperative laboratory parameters and postoperative laboratory parameters after OLT are shown in Table 2. Basal plasma syndecan-1 concentration obtained under stable anesthetic conditions after induction of anesthesia before surgical incision (TP 0) was significantly higher in liver transplant recipients than in healthy volunteers (Table 3). Furthermore, plasma levels of syndecan-1 changed significantly during and after OLT (analysis of variance, P < 0.001). There was a significant increase in plasma levels of syndecan-1 after reperfusion (TP 2) compared to basal plasma concentration (TP 0) and clamping of the inferior vena cava (TP 1) (post hoc Student-Newman-Keuls test, P < 0.05). This significant increase persisted at the end of surgery (TP 3) and 24 hours after reperfusion (TP 4) (Figure 1). We found a significant correlation between increased plasma syndecan-1 levels after reperfusion (TP 2) and aspartate aminotransferase levels on day 1 (P = 0.01; Pearson correlation coefficient r = 0.45) and day 2 (P = 0.02; Pearson correlation coefficient r = 0.4) after OLT but not on day 3 (P = 0.3; Pearson correlation coefficient r = 0.3). Mean cross-clamping time of the inferior vena cava (warm ischemia time) took 76 minutes (range, 50–100 minutes); the average duration of cold ischemia time was 425.7 minutes (range, 200–880 minutes). There was no significant correlation between warm (P = 0.4; Pearson correlation coefficient r = 0.15) or cold ischemia time (P = 0.6; Pearson correlation coefficient r = 0.1) and an increase in plasma syndecan-1 levels during and after OLT. The mean ICU and hospital length of stay were 15 ± 16 days (range, 2–70 days) and 30 ± 29 days (range, 9–148 days), respectively. We found a significant correlation between ICU and hospital length of stay (P < 0. 0001; Pearson correlation coefficient 0.84). Twenty-eight patients (93.4%) survived 28 days after OLT. One patient died due to hemorrhage and consecutive multiorgan failure, the other due to sepsis. Ninety-day and 1-year survival rates were 90% and 84%, respectively. One patient died in septic shock after
TABLE 1.
Demographic data and indications for liver transplantation Demographic data
Age, y Sex (male/female) Body surface area, m2 Child-Pugh score (A/B/C) MELD score Etiology of liver disease: Alcohol Viral Alcohol and hepatocellular carcinoma Viral and hepatocellular carcinoma Cholestatic liver disease Wilson disease Cryptogenic cirrhosis Autoimmune cirrhosis Other
54 ± 11 21 (70 %) / 9 (30 %) 1.9 ± 0.2 1 / 13 / 16 16 ± 6.9 6 (20 %) 8 (26 %) 1 (3 %) 6 (20 %) 2 (7 %) 3 (10 %) 1 (3 %) 1 (3 %) 2 (7 %)
Results are presented as mean ± standard deviation or number and percentage.
www.transplantjournal.com
re-OLT because of hepatic artery thrombosis on day 42, the other 2 patients secondary to sepsis. Plasma syndecan-1 levels after reperfusion (TP 2) were significantly lower in patients who survived 28 days (survivors 300 ± 108 ng/mL vs nonsurvivors 483 ± 50 ng/mL; P = 0.02) and 90 days (survivors 296 ± 108 ng/mL vs nonsurvivors 457 ± 56 ng/mL; P = 0.01) after OLT than in those who died. There was no significant difference in plasma syndecan-1 levels at TP 2 between patients who survived 1 year after OLT and those who died (P = 0.06). Pretransplant syndecan-1 levels were not significantly different in respect of survival after OLT. Fifteen liver graft recipients developed AKI stage 2 (n = 2) or 3 (n = 12) based on the AKIN criteria, and 10 of them were supported with continuous renal replacement therapy (CRRT). The mean duration of CRRT was 5 ± 10 days (range, 0–40 days). Furthermore, 3 of these patients underwent subsequent 1 to 5 treatments on intermittent hemodialysis before their renal function recovered. Three graft recipients developed AKI stage 1, and 12 graft recipients had no AKI, none of these required CRRT. Patients with AKI stage 2 or 3 had significantly higher plasma levels of syndecan-1 during the entire study period, compared to patients with AKI stage 0 or 1 (Table 3). We found a significant negative correlation between plasma syndecan-1 levels before surgery (TP 0) and preoperative estimated glomerular filtration rate (78.7 ± 29.3 mL/minute) with the Modification of Diet in Renal Disease formula (P = 0.002; Pearson correlation coefficient r = − 0.53). Although not significant, the mean pretransplant serum creatinine in patients who developed AKI stage 0 or 1 was lower than in patients who developed AKI 2 or 3 (0.96 ± 0.38 mg/dL vs 1.27 ± 0.49 mg/dL, respectively, P = 0.06). However, there was no significant difference in the increase of plasma levels of syndecan-1 from TP 1 to TP 2 between the patients with AKI stage 0 or 1 and patients with AKI stage 2 or 3 (unpaired t test, P = 0.19). Similarly, the mean pretransplant model of endstage liver disease score in patients who developed AKI stage 0 or 1 was lower than that in patients who developed AKI 2 or 3 (13.6 ± 6.2 vs 18.3 ± 7.0, respectively, P = 0.06). In the ROC curve analysis, syndecan-1 at TP 2 and δ syndecan-1 (TP 2 minus TP 0) exhibited an AUC of 0.76 (95% confidence interval, 0.57-0.89; P = 0.005) and 0.64 (95% confidence interval, 0.45-0.80; P = 0.17), respectively, for the development of AKI stage 2 or 3 within 48 hours after graft reperfusion. The sensitivity and specificity for syndecan-1 at TP 2 with optimal cutoff value greater than 379 ng/mL were 60% and 100%, respectively. The sensitivity and specificity for δ syndecan-1 with optimal cutoff value greater than 314 ng/mL were 40 % and 93 %, respectively. General anesthesia was maintained with volatile anesthetics (sevoflurane, n = 7; desflurane, n = 5) or with propofol-based total intravenous anesthesia (n = 18). The type of general anesthesia had no significant effect on plasma levels of syndecan-1 (unpaired t test; P = 0.28). There was no difference between patients anesthetized with sevoflurane and desflurane in regard to syndecan-1 levels. DISCUSSION This study is the first to describe the perioperative changes of plasma syndecan-1 as a marker of endothelial glycocalyx damage during human OLT. Notably, the baseline plasma
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2121
Schiefer et al
© 2015 Wolters Kluwer
TABLE 2.
Baseline and postoperative laboratory parameters before and after liver transplantation in 30 patients with end-stage liver disease Parameter
Sodium, mmol/L Crea, mg/dL Bilirubin, mg/dL AST, U/L ALT, U/L Hb, mg/dL WBC, G/L Plt, G/L PT (%)
pre-OLT
POD 1
POD 2
POD 3
POD 7
136 ± 4 1.1 ± 0.4 4.7 ± 6 123 ± 157 112 ± 265 11.2 ± 1.6 5.7 ± 2.5 98.6 ± 54.5 56 ± 20
138 ± 4 1.26 ± 0.4 5.0 ± 3.1 1589 ± 1198a 849 ± 526 9.5 ± 1.2 11.4 ± 4.4 90.0 ± 74.4 46 ± 13
139 ± 4 1.63 ± 0.7 5.9 ± 4.3 1329 ± 1506* 776 ± 581 9.2 ± 1.4 10.9 ± 4.9 63.0 ± 64.1 43 ± 12
138 ± 4 1.69 ± 0.9 5.1 ± 3.6 644 ± 723 648 ± 548 8.8 ± 1.1 6.9 ± 3.9 42.3 ± 30.9 51 ± 14
137 ± 4 1.26 ± 0.5 5.3 ± 5.0 278 ± 1057 304 ± 545 9.3 ± 1.2 8.4 ± 6.1 61.1 ± 37.2 69 ± 12
a
A significant correlation between AST and plasma syndecan-1 levels after reperfusion due to IRI (P < 0.05). Results are presented as mean ± standard deviation. POD, postoperative day; Crea, creatinine; AST, aspartate aminotransferase; ALT, alanine aminotransferase; Hb, hemoglobin; WBC, white blood cells; Plt, platelet count; PT, prothrombin time.
syndecan-1 before surgery was significantly higher (7-fold) in liver transplant recipients compared to healthy volunteers, reflecting the pathophysiological alteration of glycocalyx in ESLD patients. Although no data on syndecan-1 in ESLD patients were previously published, the baseline syndecan-1 levels in our transplant recipients before OLT were comparable to the plasma syndecan-1 concentrations recently described after major abdominal surgery.9 The pronounced shedding of syndecan-1 before OLT may result from the derangement of inflammatory response seen in patients with liver cirrhosis with consequent overexpression of various proinflammatory and anti-inflammatory cytokines.18 Furthermore, perioperative proinflammatory cytokine levels and permeability of sinusoids in cirrhotic livers were shown to correlate with the severity of the liver disease necessitating OLT.11,18,19 Among other possible pathophysiological mechanisms, bacterial translocation, often occurring in ESLD, is an important trigger of systemic inflammation and may thus potentiate vascular dysfunction in ESLD.20 Bacterial lipopolysaccharide and tumor necrosis factor-α were both shown to cause glycocalyx shedding.21,22 In line with our results in ESLD patients, compositional and dimensional changes of glycocalyx in sublingual microvasculature were recently described in patients with end-stage renal disease.15 Reduced renal function was strongly associated with low ESL dimension. Although not yet reported in human OLT, the significant increase of plasma syndecan-1 after reperfusion in our patients is in accordance with data from animal and human
studies investigating glycocalyx shedding after IRI. Major degradation of glycocalyx was described after 20 minutes in an isolated heart model and after 60 minutes of warm ischemia followed by reperfusion in a less sensitive bowel model.7 Human studies confirmed these results and showed a concomitant increase of glycocalyx degradation products with reperfusion.8,23 Notably, plasma syndecan-1 after reperfusion in our collective reached concentrations as high as previously described in septic patients,9 and was 29.4-fold higher than in healthy controls. Furthermore, levels of proinflammatory markers, such as interleukin-6, were correlated with syndecan-1 levels in septic patients.9 This is in line with our previous results in patients with ESLD undergoing OLT without venovenous bypass showing significant increase of interleukin-6 after reperfusion.24 There was also a significant positive correlation between syndecan-1 after reperfusion and serum aspartate aminotransferase on the first 2 postoperative days, reflecting the damage caused by IRI. Although transaminase activity levels represent a well-established parameter for the estimation of hepatic injury after OLT, their increased levels alone are not predictive of outcome or incidence of early allograft dysfunction. Because of a small sample size, we cannot conclude about the predictive value of syndecan-1 alone on early mortality after OLT, but plasma syndecan-1 levels after reperfusion were significantly lower in patients who survived 28 and 90 days after OLT than in those who died. Because pretransplant syndecan-1 levels were not significantly different with respect of survival after OLT, factors influencing integrity of glycocalyx, such as
TABLE 3.
Plasma levels of syndecan-1 (ng/mL) in human liver transplantation
Control TP 0 TP 1 TP 2 TP 3 TP 4 δ syndecan-1
All patients n = 30
AKIN 0 – 1 n = 15
AKIN 2 – 3 n = 15
p
10.6 ± 9.4 74.3 ± 59.9 83.5 ± 53.3 312.6 ± 114.8 291.8 ± 117.5 279.7 ± 138.3 238.3 ± 122.5
49.3 ± 50.1 60.4 ± 52.9 257.7 ± 82.1 248.1 ± 97.6 202.8 ± 88.1 208.4 ± 86.8
99.1 ± 59.8 106.7 ± 44.1 367.4 ± 118.9 335.4 ± 122.4 351.5 ± 140.0 268.2 ± 146.9
0.018 0.015 0.006 0.039 0.002 0.19
δ syndecan-1 represents difference between TP 0 and TP 2, Results are presented as mean ± standard deviation. P value applies for comparison of plasma syndecan-1 concentration in patients with AKI stage 0 or 1 versus AKI stage 2 and 3.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
2122
Transplantation
■
October 2015
■
Volume 99
■
Number 10
Figure 1. Measurements of circulatory endothelial surface layer damage marker syndecan-1 in patients with end-stage liver disease undergoing liver transplantation. *TP 0 and TP 1 versus healthy volunteers (P < 0.05); **TP 2, TP 3, and TP 4 versus TP 0 (P < 0.05).
IRI, quality of graft and ischemia time, could have potentially influenced the outcome of our patients. Acute kidney injury after OLT is a recognized independent risk factor for morbidity, mortality, reduced graft survival, longer hospital stay, and higher costs.17,25 Recent studies using a consensus definition based on the AKIN showed that when AKI is defined as at least a doubling of serum creatinine levels or the need for RRT (AKIN stages 2 or 3), the incidence of AKI in OLT is up to approximately 48%.17 This is in agreement with our results because we found AKI stage 2 or 3 in 50% of our patients and 33% of these received CRRT. However, serum creatinine is slow and insensitive because it does not immediately reflect the dynamic changes in renal function and therefore does not alert the clinician of impeding renal injury. In our study, AUC of the ROC analysis of plasma syndecan-1 and increase of syndecane-1 after reperfusion to predict AKI (stage 2 or stage 3) was 0.75 and 0.64, respectively. However, because of the low sensitivity, plasma syndecan-1 is not helpful as an alternative marker of de novo AKI in human OLT. In a study from Wagener et al,25 urinary neutrophil gelatinase-associated lipocalin-tourine creatinine ratio demonstrated encouraging results with much better discrimination power to predict AKI stage 3 after OLT. According to our results, patients who developed AKI stage 2 or 3 had significantly higher plasma syndecan-1 compared to graft recipients with AKI stage 0 or 1 before OLT, and it remained significantly higher throughout the whole study. Furthermore, a significant negative correlation between estimated glomerular filtration rate and syndecan-1 before surgery was found in our patients. Serum creatinine was also slightly higher before OLT (with a mean difference of 0.3 mg/dL) in patients who developed AKI stage 2 or 3 compared to those with AKI stage 0 or 1 after OLT. Bearing in mind that all transplant recipients in this study had, compared to healthy volunteers, significantly higher plasma levels of syndecan-1 before OLT, these results suggest that also slightly elevated or nearly normal serum creatinine (due to malnutrition and reduced muscle mass) in combination with pre-OLT increased syndecan-1, reflects a higher grade of chronic inflammatory injury of ESLD predisposing to AKI and higher mortality. Reduced renal clearance of syndecan-1 could also be involved in its increased plasma
www.transplantjournal.com
concentration, but its large molecular weight excludes a direct renal excretion. Although exact clearance of syndecan-1 is still unknown, low levels of syndecan-1 were shown in patients with interstitial fibrosis/tubular atrophy after kidney transplantation and reduced renal function.15 Our results are in line with a previous study by Dane et al15 that showed a strong association of reduced renal function with low endothelial surface dimensions in patients with end-stage renal disease. After successful kidney transplantation, the ESL of the sublingual microvasculature, measured by a novel noninvasive sidestream-darkfield imaging model, was indistinguishable from healthy control. Similarly, successful OLT often leads to improvement of reduced renal function seen in ESLD or even reversal of hepatorenal syndrome. Although 10 patients in our study were supported with CRRT and 3 patients later with intermittent hemodialysis, none of them required extracorporeal renal support at the end of follow-up. The negative impact of IRI on the integrity of glycocalyx contributing to endothelial dysfunction makes its protection and/or restoration a valuable therapeutic goal, and preconditioning and postconditioning with various pharmacological agents in setting of IRI was recently studied.3,6,14,26 Lucchinetti et al14 reported a markedly improved postocclusive hyperaemic reaction and attenuated activation of leukocytes with peri-ischemic inhalation of sevoflurane, targeted at 0.5 to 1 vol% end-tidal concentrations, after 15 minutes of forearm ischemia followed by reperfusion in 5 healthy volunteers. The targeted end-tidal concentration of sevoflurane in this study was rather low, but approximates the end-tidal concentration of sevoflurane used in our study during cross-clamping of the inferior vena cava and thus warm ischemia and early reperfusion phase of human OLT. Furthermore, in an isolated guinea pig heart model, inhalation of 1 minimal alveolar concentration sevoflurane protected the endothelial glycocalyx from IRI-induced degradation, with both preconditioning and rapid postconditioning being successful.6 The postconditioning (during reperfusion period) in this study mimics the clinical setting of human OLT because in case of volatile anesthetics, the liver graft is exposed to sevoflurane first during reperfusion. Contrary to these experimental studies, we observed no differences in plasma syndecan-1 at any time during the OLT, regardless of the use of volatile anesthetics or propofol-based anesthesia. This may be due to cold ischemia time that was not part of the experimental studies but is currently unavoidable in real-life clinical transplantation and considerably longer warm ischemia than in the studies mentioned above. Because we did not perform a power analysis on the effect of volatile anesthetics on glycocalyx shedding, we cannot exclude any beneficial effects of volatile anesthetics in a larger study population, but the previous studies found significant differences in similar or even smaller-sized study populations. In line with the lack of protective effects of volatile anesthetics on glycocalyx shedding in our study, the beneficial effects of corticosteroids, assessed previously by reduced transudate formation, leukocyte adhesion, and decreased shedding of syndecan-1 after IRI in animal models, were again in contrast with our results.12,27 According to our immunosuppressive protocol, 40 mg of dexamethasone was intravenously injected before cross clamping of the inferior vena cava. There was no steroid in the preservation or flushing solution because it was part of the experimental protocol
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Schiefer et al
© 2015 Wolters Kluwer
in previously mentioned successful animal models.12,27 Our data are in line with the results reported by Biancofiore et al,18 where the administration of cortisone during OLT as a part of immunosuppressive therapy did not change the inflammatory pattern nor affected the cytokine balance by shifting it toward inflammation or over-immunosuppression. Based on these major discrepancies, no conclusion can be drawn on the potentially beneficial effect of steroids on glycocalyx shedding in clinical settings without adding steroids to preservation solutions. However not unexpected, 40 mg of dexamethasone intravenously before cross-clamping of the caval vein in human OLT definitively does not prevent glycocalyx from pronounced shedding. Similar to this model, augmenting histidine-tryptophan-ketoglutarate solutions with human albumin improved endothelial integrity and heart performance after 4 hours of cold ischemia of guinea pig hearts.28 Although we routinely flushed all liver grafts with 20% human albumin before reperfusion for at least 20 minutes, massive syndecan-1 increase has not been attenuated. The different timing of albumin administration most probably hinders the translation of beneficial effects seen under experimental conditions into clinical setting and makes further clinical investigations necessary. Our study has several limitations. Because we collected our blood samples from the radial artery, only systemic changes in syndecan-1 could be investigated. The flushing of liver grafts with human albumin before reperfusion could wash out some syndecan-1 molecules. The contribution of OLT as major abdominal surgery additionally to IRI-induced glycocalyx shedding cannot be distinguished under these circumstances. Furthermore, cross-clamping of the inferior vena cava with consecutive renal and splanchnic venous congestion in OLT, without the use of venovenous bypass, is at least in part responsible for the pronounced syndecan-1 increase seen after reperfusion. In conclusion, plasma syndecan-1 concentration as a marker of glycocalyx shedding in patients with ESLD is significantly increased when compared to healthy population. Ischemia-reperfusion injury during OLT further exacerbates glycocalyx shedding. Despite a higher incidence of AKI and the need for CRRT in patients with higher syndecan-1 concentration, syndecan-1 is not helpful as an alternative biomarker to predict de novo AKI. Contrary to previous experimental and animal studies, rapid postconditioning with volatile anesthetics did not attenuate glycocalyx shedding under clinical condition of human OLT. REFERENCES 1. Seal JB, Gewertz BL. Vascular dysfunction in ischemia-reperfusion injury. Ann Vasc Surg. 2005;19:572–584. 2. Zhai Y, Petrowsky H, Hong JC, Busuttil RW, Kupiec-Weglinski JW. Ischaemia-reperfusion injury in liver transplantation—from bench to bedside. Nat Rev Gastroenterol Hepatol. 2013;10:79–89. 3. Becker BF, Chappell D, Bruegger D, Annecke T, Jacob M. Therapeutic strategies targeting the endothelial glycocalyx: acute deficits, but great potential. Cardiovasc Res. 2010;87:300–310. 4. Salmon AHJ, Satchell S. Endothelial glycocalyx dysfunction in disease: albuminuria and increased microvascular permeability. J Pathol. 2012; 226:562–574. 5. Weinbaum S, Tarbell JM, Damiano ER. The structure and function of the endothelial glycocalyx layer. Annu Rev Biomed Eng. 2007;9:121–167.
2123
6. Annecke T, Chappell D, Chen C, et al. Sevoflurane preserves the endothelial glycocalyx against ischaemia-reperfusion injury. Br J Anaesth. 2010;104:414–421. 7. Mulivor AW, Lipowsky HH. Inflammation- and ischemia-induced shedding of vebular glycocalyx. Am J Physiol Heart Circ Physiol. 2004;286: H1672–H1680. 8. Rehm M, Bruegger D, Christ F, Conzen P, Thiel M, Jacob M. Shedding of endothelial glycocalyx in patients undergoing major vascular surgery with global and regional ischemia. Circulation. 2007;116:1896–1906. 9. Steppan J, Hofer S, Funke B, et al. Sepsis and major abdominal surgery lead to flaking of the endothelial glycocalix. J Surg Res. 2011; 165:136–141. 10. Nelson A, Beckerstedt I, Schmidtchen A, Ljunggren L, Bodelsson M. Increased levels of glycosaminoglycans during septic shock: relation to mortality and the antimicrobial actions of plasma. Shock. 2008; 30:623–627. 11. Koh HJ, Ryu KH, Cho ML, Heo YJ, Lee J. Factors influencing the concentration of cytokines during liver transplantation. Transplant Proc. 2010;42:3617–3619. 12. Chappell D, Dörfler N, Jacob M, et al. Glycocalyx protection reduces leucocyte adhesion after ischemia/reperfusion. Shock. 2010; 34: 133–139. 13. Chappell D, Heindl B, Jacob M, et al. Sevoflurane reduces leukocyte and platelet adhesion after ischemia-reperfusion by protecting the endothelial glycocalyx. Anesthesiology. 2011;115:483–491. 14. Lucchinetti E, Ambrosio S, Aguirre J, et al. Sevoflurane inhalation at sedative concentrations provides endothelial protection against ischemia-reperfusion injury in humans. Anesthesiology. 2007;106: 262–268. 15. Dane MJC, Khairoun M, Lee DH, van der Berg BM, Eskens BJM, Boels MGS. Association of kidney function with changes in the endothelial surface layer. Clin J Am Soc Nephrol. 2014;4:698–704. 16. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130:461–470. 17. Razonable RR, Findlay JY, O'Riordan A, et al. Critical care issues in patients after liver transplantation. Liver Transpl. 2011;17:511–527. 18. Biancofiore G, Bindi L, Miccoli M, et al. Balance of pro- and anti-inflammatory cytokines in cirrhotic patients undergoing liver transplantation. Transplant Immunol. 2013;28:193–197. 19. Villeneuve JP, Daganais M, Hiet PM, Lapointe R, Marleau D. The hepatic microcirculation in the isolated perfused human liver. Hepatology. 1996;23:24–31. 20. Metha G, Gustot T, Mookerjee RP, et al. Inflammation and portal hypertension—the undiscovered country. J Hepatol. 2014 http://dx.doi. org/10.1010/j.jhep.2014.03.014. 21. Henry CB, Duling BR. TNF-alpha increases entry of macromolecules into luminal endothelial cell glycocalyx. Am J Physiol Heart Circ Physiol. 2000;279:H82815–H82823. 22. Nieuwdrop M, Meuwese MC, Mooij HL, et al. Tumor necrosis factor-alpha inhibition protects against endotoxin-induced endothelial glycocalyx perturbation. Atherosclerosis. 2009;202:296–303. 23. Bruegger D, Rehm M, Abicht J, et al. Shedding of the endothelial glycocalyx during cardiac surgery: on-pump versus off-pump coronary artery bypass graft surgery. J Thorac Cardiovasc Surg. 2009; 138:1445–1447. 24. Faybik P, Hetz H, Krenn CG, Baker A, Berlakovich GA, Steltzer H. Perioperative cytokines during orthotopic liver transplantation without venovenous bypass. Transplant Proc. 2003;35:3019–3021. 25. Wagener G, Minhaz M, Mattis FA, Kim M, Emond JC, Lee HT. Urinary neuthrophil gelatinase-associated lipocalin as a marker of acute kidney injury after orthotopic liver transplantation. Nephrol Dial Transplant. 2011;26:1717–1723. 26. Smit KF, Oei GT, Brevoord D, Stroes ES, Nieuwland R, Schlack WS. Helium induces preconditioning in human endothelium in vivo. Anesthesiology. 2013;118:95–104. 27. Chappell D, Jacob M, Hofmann-Kiefer K, Bruegger D, Rehm M, Conzen P. Hydrocortisone preserves the vascular barrier by protecting the endothelial glycocalyx. Anesthesiology. 2007;107:776–784. 28. Jacob M, Mehringer PO, Chappell D, et al. Albumin augmentation improves condition of guinea pig hearts after 4 hr of cold ischemia. Transplantation. 2009;87:956–965.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
