Abstract
Early allograft dysfunction (EAD) is associated with graft failure and mortality after living donor liver transplantation (LDLT). In this study, we report biomarkers superior to other conventional clinical markers in the prediction of EAD and all-cause in-hospital mortality in LDLT patient cohort. Blood samples of living donor liver transplant recipients were collected on postoperative day 1 and analyzed by liquid chromatography coupled with mass spectrometry (LC-MS). Significant metabolites associated with the prediction of EAD were identified using orthogonal projection to latent structures-discriminant analysis (OPLS-DA). A few lipids, more specifically, lysoPC (16:0), PC (18:0/20:5), betaine and palmitic acid (C16:0) were found to effectively differentiate EAD from non-EAD on postoperative day 1. A combination of these four metabolites showed an AUC of 0.821, which was further improved to 0.846 by the addition of a clinical parameter, total bilirubin. The panel exhibits a high prognostic accuracy in prediction of all-cause in-hospital mortality and mortality within 7 postoperative days with AUCs of 0.843 and 0.954. These results show the combination of metabolomics-derived biomarkers and clinical parameters demonstrates the power of panels in diagnostic and prognostic evaluation of LDLT.
Keywords: Lipidomics, early allograft dysfunction, phosphatidylcholines, lysophosphatidylcholines, betaine, living donor liver transplantation
Introduction
Liver transplantation is a life-saving treatment for patients with end stage liver disease and an alternative therapy for patients with hepatocellular carcinoma (HCC); however, a shortage of liver donors has always been a major limitation for liver transplantation. The concept of living donor liver transplantation (LDLT) has recently become widely accepted with a one-year survival rate approximately 90%, better than that of deceased donor liver transplantation (DDLT) [1,2]. That said, primary graft dysfunction, a syndrome encompassing the milder, reversible form of early allograft dysfunction (EAD) to the more severe, irreversible form of primary non-function (PNF) can occur post-transplantation [3,4]. Although no consensus in EAD definition is reached, clinically, EAD is recognized frequently based upon deranged liver function, such as elevated serum alanine aminotransferase (ALT), aspartate aminotransferase (AST) and bilirubin levels and the prolonged international normalized ratio (INR) within the first seven days postoperatively [5]. EAD develops at an incidence ranging from 5.2% to 36.3% [4] with an EAD-associated mortality reaching 18.8% [5,6]. EAD is attributed by a number of factors including hepatic steatosis in donors [7], diagnosis of hepatocellular carcinoma and dialysis at transplant in recipients [8], intraoperative blood loss requiring blood transfusion [9] and long ischemia time [3]. Although age has long been recognized as a risk factor for the development of EAD, recent studies have shown more liberal results for both older donors and recipients [10,11]. The literature has shown that disturbed lipid homeostasis may be related to the functional status of liver [12]. In the setting of EAD, our team has previously utilized 1H-nuclear magnetic resonance (1H-NMR) spectroscopy and liquid chromatography coupled with mass spectrometry (LC-MS) in identifying lipidomic profiles in association with EAD. Previously, we have demonstrated that on postoperative day 7, the changes in lipids serve as prognostic factors for short term outcomes [13]. In this study, we aim to identify novel metabolites that may predict the event of EAD on postoperative day 1 in the hope to initiate timely treatment after LDLT.
Materials and methods
Patients
The study received prior approval from the Institutional Review Board of Chang Gung Memorial Hospital (IRB 1805230007) and registered under The Australian New Zealand Clinical Trials Registry (ACTRN12619000386134). All subjects gave their informed consent for inclusion before they participated in the study. After applying exclusion criteria, including concurrent septic shock status, a preoperative measured pulmonary wedge pressure greater than 35 mmHg, or refusal to give informed consent, 74 recipients undergoing LDLT between May 2015 and February 2018 were recruited to the study.
General anesthesia was conducted for LDLT [14]. Within 7 days postoperatively, the primary allograft function was assessed based on Olthoff’s criteria: international normalized ratio (INR) ≥ 1.6 or bilirubin ≥ 10 mg/dL on postoperative day 7 or aspartate (AST) or alanine (ALT) aminotransferaess >2000 IU/mL within the first 7 postoperative days [5,15]. Grafts meeting one or more of the aforementioned criteria was characterized as EAD. The short-term mortality was also assessed as secondary outcomes.
Blood samples
A peripherally indwelling arterial catheter was inserted in the recipient perioperatively, from which the blood samples were collected on postoperative day 1. The blood was centrifuged immediately to obtain plasma. The biochemistry was sent to the central clinical laboratory for analysis.
LC-MS-lipidomic analysis
The plasma sample and precooling isopropanol (IPA) was mixed. The mixture was vortexed for 60 seconds and then stand on ice for 30 min. After incubation, the mixture was centrifuged at 12,000 rpm at 4°C for 30 min. The supernatant was transferred to glass tube. Finally, the clear supernatant was collected for LC-MS analysis [16].
Statistical analysis
The continuous variables data were presented as the mean ± standard deviation (SD). The categorical data were presented as frequencies and compared using the chi-square test or the Fisher’s exact test. To identify an independent predictor of postoperative EAD, linear logistic regression analysis was performed. A receiver operator characteristic curve (ROC) was created to compare the predictive accuracy of the identified variables. A p-value <0.05 was considered to be significantly different. Analyses were performed in SAS 9.4 and R 3.3.2 [17]. Several software programs were utilized to analyze metabolomic data. The orthogonal projection to latent structures-discriminant analysis (OPLS-DA) model was performed. Each variable in the model was given a variable importance in the projection (VIP) value.
Results
Patient characteristics
74 patients received LDLT and 22 developed EAD on postoperative day 7 whereas the other 51 had uneventful recovery. Table 1 summarized the preoperative demographics, intraoperative parameters and postoperative secondary outcomes. EAD and non-EAD groups had a mean age of 53.33 ± 9.03 and 56.45 ± 7.10 years, respectively. The incidence of EAD was 29.7%. Of the pre-operative parameters, no statistical significance was observed for patient age, height, weight, body mass index (BMI), gender and the etiologies of liver transplantations except for the preoperative MELD scores, which were statistically higher in the EAD group (21.22 ± 11.19 versus 16.41 ± 8.37, p-value 0.0438). Although donor-to-recipient gender mismatch and ABO incompatibility appeared to be risk factors for poor graft survival after liver transplantation [18,19], no statistical significance of either in the event of EAD was recognized in our study. Intraoperatively, the EAD group, compared to non-EAD group, had slightly shorter cold ischemia time (43.39 ± 30.07 versus 51.49 ± 32.82 minutes, respectively) and warm ischemia time (45.09 ± 26.66 versus 50.02 ± 39.31 minutes, respectively) with no statistical significance. Clinically, the EAD group appeared to endure more blood loss than non-EAD group (2547.83 ± 1806.71 versus 1712.16 ± 1552.97 mL, respectively, p-value 0.0455) and required more red blood cell transfusions intraoperatively. Among the 74 recipients, the difference in mortality within 7 postoperative days between EAD and non-EAD groups within 7 days was not computed as no mortality was observed in the non-EAD group. More mortality was observed in the EAD group within 30 days, 60 days and 90 days postoperatively with statistical significance (p-value 0.0380, 0.0090 and 0.0070, respectively). Eighteen patients had a graft recipient weight ratio (GRWR) less than 0.8%, satisfying the definition for small for graft syndrome (SFGS) [20,21].
Table 1.
A summary of the clinical parameters of LDLT recipients
| EAD (N = 23) | Non-EAD (N = 51) | p-value | |
|---|---|---|---|
| PRE-OPERATIVE PARAMETERS | |||
| Age | 52.48 ± 10.20 | 56.45 ± 7.10 | 0.1004 |
| Height | 1.65 ± 0.09 | 1.62 ± 0.08 | 0.1383 |
| Weight | 68.74 ± 13.59 | 66.77 ± 11.34 | 0.5178 |
| BMI | 25.18 ± 3.77 | 25.39 ± 3.33 | 0.8153 |
| MELD | 21.22 ± 11.19 | 16.41 ± 8.37 | 0.0438 |
| Gender (F/M) | 6/17 | 25/26 | 0.0657 |
| Gender Mismatch | 16 | 23 | 0.0520 |
| Female to Male | 11 | 15 | 0.1281 |
| Blood Type | 0.9577 | ||
| A | 8 | 18 | |
| B | 5 | 11 | |
| O | 9 | 20 | |
| AB | 1 | 2 | |
| ABO Incompatibility | 3 | 8 | 0.7712 |
| Etiology | |||
| HBV | 11 | 23 | 0.8303 |
| HCV | 6 | 18 | 0.4405 |
| Alcoholism | 7 | 14 | 0.7956 |
| HCC | 6 | 23 | 0.1244 |
| INTRA-OPERATIVE PARAMETERS | |||
| Blood Loss (mL) | 2547.83 ± 1806.71 | 1712.16 ± 1552.97 | 0.0455 |
| RBC (units) | 11.61 ± 7.83 | 7.57 ± 7.16 | 0.0324 |
| FFP (units) | 15.48 ± 10.72 | 10.82 ± 10.34 | 0.0806 |
| Platelets (units) | 9.39 ± 8.83 | 8.24 ± 9.14 | 0.6124 |
| Graft Size (mg) | 626.78 ± 127.62 | 637.25 ± 140.63 | 0.7614 |
| GRWR (%) | 0.94 ± 0.18 | 0.98 ± 0.29 | 0.4050 |
| CIT (minutes) | 43.39 ± 30.07 | 51.49 ± 32.82 | 0.3171 |
| WIT (minutes) | 45.09 ± 26.66 | 50.02 ± 39.31 | 0.5307 |
| POST-OPERATIVE OUTCOMES | |||
| Mortality within 7 days | 2 | 0 | N/A |
| Mortality within 30 days | 5 | 1 | 0.0380 |
| Mortality within 60 days | 7 | 1 | 0.0090 |
| Mortality within 90 days | 8 | 2 | 0.0070 |
| In hospital Mortality | 8 | 3 | 0.0117 |
Abbreviations: EAD, early allograft dysfunction; BMI, body mass index; MELD, model for end-stage liver disease; HBV, hepatitis B virus; HCV, hepatitis C virus; HCC, hepatocellular carcinoma; PRBC, packed red blood cell; FFP, fresh frozen plasma; GRWR, graft recipient weight ratio; CIT, cold ischemia time; WIT, warm ischemia time; ICU, intensive care unit.
Preoperatively, the majority of lab data showed no difference except for glomerular filtration rate (GFR), total protein and bilirubin levels (Table 2). On postoperative day 1 (T2), EAD group showed significantly worse coagulation, renal and liver function and electrolyte imbalance. Similarly, on postoperative day 7 (T3), EAD group demonstrated significantly poorer coagulation profiles and liver function than non-EAD group, in support of the diagnosis of EAD.
Table 2.
Biochemical data preoperatively (T1) on postoperative day 1 (T2) and day 7 (T3)
| T1 | T2 | T3 | |||||||
|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
|||||||
| EAD (N = 23) | Non-EAD (N = 51) | p-value | EAD (N = 23) | Non-EAD (N = 51) | p-value | EAD (N = 23) | Non-EAD (N = 51) | p-value | |
| Hemoglobin (g/dL) | 9.72 ± 2.33 | 10.73 ± 2.40 | 0.0971 | 10.88 ± 3.02 | 10.91 ± 3.17 | 0.9661 | 11.1 ± 2.2 | 10.3 ± 1.3 | 0.1065 |
| Hematocrit (%) | 28.91 ± 6.90 | 31.64 ± 6.75 | 0.1142 | 31.35 ± 8.42 | 31.14 ± 6.59 | 0.9067 | 32.1 ± 5.7 | 30.0 ± 3.9 | 0.14603 |
| Platelet (1000/dL) | 77.35 ± 58.34 | 84.44 ± 43.94 | 0.5644 | 53.96 ± 37.78 | 68.06 ± 38.60 | 0.1476 | 47.5 ± 30.3 | 70.0 ± 37.0 | 0.0163 |
| INR (seconds) | 1.79 ± 0.65 | 1.63 ± 0.68 | 0.3448 | 2.11 ± 0.59 | 1.71 ± 0.31 | 0.0053 | 1.5 ± 0.4 | 1.2 ± 0.1 | 0.0016 |
| BUN (mg/dL) | 32.88 ± 33.86 | 17.80 ± 17.41 | 0.0535 | 34.75 ± 25.33 | 24.63 ± 15.29 | 0.0861 | 56.7 ± 44.1 | 23.8 ± 13.4 | 0.0029 |
| Creatinine (mg/dL) | 1.69 ± 1.75 | 1.02 ± 1.33 | 0.0732 | 1.95 ± 1.23 | 1.13 ± 0.76 | 0.0056 | 1.9 ± 1.4 | 0.8 ± 0.5 | 0.0008 |
| GFR (mL/mL/m/1.73 min/1.73 m2) | 79.74 ± 51.46 | 112.25 ± 51.34 | 0.0140 | 47.36 ± 35.01 | 86.43 ± 49.78 | 0.0011 | 64.6 ± 64.3 | 116.7 ± 45.3 | 0.0021 |
| Total protein (g/dL) | 6.11 ± 1.17 | 6.62 ± 0.88 | 0.0444 | 4.17 ± 0.85 | 4.74 ± 0.65 | 0.0026 | 4.1 ± 0.7 | 4.4 ± 0.5 | 0.0476 |
| Albumin (g/dL) | 3.03 ± 0.64 | 3.09 ± 0.71 | 0.7544 | 2.36 ± 0.38 | 2.55 ± 0.41 | 0.0579 | 2.5 ± 0.5 | 2.7 ± 0.4 | 0.2575 |
| Na (mEq/L) | 140.22 ± 5.66 | 138.43 ± 3.92 | 0.1794 | 142.17 ± 5.90 | 139.00 ± 4.31 | 0.0111 | 130.6 ± 27.5 | 137.4 ± 3.4 | 0.2698 |
| K (mEq/L) | 3.58 ± 0.54 | 3.75 ± 0.55 | 0.2230 | 3.71 ± 0.56 | 3.71 ± 0.57 | 0.9842 | 3.8 ± 0.7 | 3.6 ± 0.5 | 0.2033 |
| Ca (mEq/L) | 8.26 ± 0.51 | 8.27 ± 0.70 | 0.9678 | 7.38 ± 0.65 | 7.48 ± 0.70 | 0.5628 | 7.9 ± 0.7 | 7.4 ± 0.9 | 0.0284 |
| Sugar | 129.70 ± 42.22 | 138.59 ± 52.99 | 0.4807 | 214.83 ± 76.18 | 282.67 ± 102.53 | 0.0059 | 184.1 ± 72.1 | 178.5 ± 55.9 | 0.7300 |
| Bilirubin total (mg/dL) | 13.17 ± 15.68 | 4.61 ± 6.43 | 0.0183 | 12.34 ± 9.25 | 4.55 ± 3.42 | 0.0006 | 12.1 ± 7.8 | 2.7 ± 2.1 | <0.0001 |
| Bilirubin direct (mg/dL) | 7.65 ± 9.81 | 2.17 ± 3.24 | 0.0152 | 6.00 ± 5.29 | 1.99 ± 1.77 | 0.0016 | 7.0 ± 4.8 | 1.5 ± 1.4 | <0.0001 |
| AST (U/L) | 103.78 ± 141.52 | 68.75 ± 51.87 | 0.2600 | 720.35 ± 923.72 | 295.69 ± 220.09 | 0.0399 | 147.2 ± 191.9 | 110.4 ± 77.0 | 0.40391119 |
| ALT (U/L) | 54.70 ± 34.37 | 38.82 ± 22.45 | 0.0513 | 764.17 ± 1059.86 | 304.41 ± 228.64 | 0.0510 | 273.7 ± 289.6 | 201.3 ± 162.9 | 0.2916 |
Abbreviations: EAD, early allograft dysfunction; INR, internationalized ratio; BUN, blood urea nitrogen; GFR, glomerular filtration rate; Na, sodium; K, potassium; Ca, calcium; LDH, lactate dehydrogenase; AST, aspartate aminotransferase; ALT, alanine aminotransferase.
Circulatory lipid profiles in recipients using LC-MS
Ultra performance liquid chromatography-time of flight mass spectrometry (UPLC-TOFMS) was performed with the plasma samples collected from EAD and non-EAD recipients on postoperative day 1. Figure 1A shows the OPLS-DA score plot for EAD (red) and non-EAD (blue) datasets collected in electrospray positive ion mode endowed with R2X = 0.793, R2Y = 0.843, Q2 = 0.664, representing the explanation, fitness and prediction power of the model. Additionally, Figure 1B shows OPLS-DA score plot for EAD (red) and non-EAD (blue) datasets collected in electrospray negative ion mode endowed with R2X = 0.666, R2Y = 0.735 and Q2 = 0.529. Heatmap analysis (Figure 1C) and OPLS-DA demonstrated a separation of metabolites distinguishing EAD from non-EAD. Twenty-nine metabolites were selected and shown in Table 3. They included betaine, free fatty acids, such as palmitic acid, linoleic acid, oleic acid, lysophosphatidylcholines (lysoPC (16:0)), and phosphatidylcholines (PCs). The levels of betaine, free fatty acids and PC (32:0 and 32:1) appeared to be elevated whereas other PC species were lowered in the EAD group as compared to non-EAD group.
Figure 1.
Metabolomic analysis of the plasma samples from EAD and non-EAD recipients. A. The OPLS-DA score plot for EAD (red) and Non-EAD (blue) datasets collected in electrospray positive ion mode (with parameters R2X = 0.793, R2Y = 0.843, Q2 = 0.664). B. The OPLS-DA score plot for EAD (red) and Non-EAD (blue) datasets collected in electrospray negative ion mode (with parameters R2X = 0.666, R2Y = 0.735, Q2 = 0.529). C. Heatmap analysis distinguishing EAD from non-EAD recipients. Abbreviations: EAD, early allograft dysfunction; OPLS-DA, orthogonal projection to latent structures-discriminant analysis.
Table 3.
A summary of metabolites distinguishing the EAD from the non-EAD group on postoperative day 1 (VIP >1 and p-value <0.05)
| ID | Adduct | EAD (N = 23) | Non-EAD (N = 51) | p-value | VIP |
|---|---|---|---|---|---|
| Betaine | M+H | 7286.22 ± 5287.27 | 4738.75 ± 1738.33 | 0.0002 | 4.57 |
| Palmitic acid (C16:0) | M-H | 458.96 ± 297.70 | 297.00 ± 231.01 | 0.0001 | 2.18 |
| Oleic acid (C18:1) | M-H | 2470.17 ± 1959.52 | 1555.71 ± 1783.92 | 0.0007 | 4.60 |
| LysoPC (16:0) | M+H | 861.00 ± 683.70 | 1454.67 ± 721.06 | <0.0001 | 2.28 |
| PC (16:0/16:1) | M+H | 4029.76 ± 1719.00 | 2850.05 ± 1237.37 | <0.0001 | 3.22 |
| PC (16:0/16:0) | M+H | 10269.97 ± 4303.20 | 6959.97 ± 2268.48 | <0.0001 | 5.99 |
| PC (16:0/20:5) | M+H | 1903.41 ± 980.02 | 2906.62 ± 1996.07 | <0.0001 | 2.50 |
| PC (16:0/20:4) | M+H | 43859.61 ± 13800.01 | 52715.85 ± 19138.16 | 0.0001 | 6.96 |
| PC (16:0/22:6) | M+H | 19410.38 ± 5771.54 | 27480.17 ± 12769.92 | <0.0001 | 7.90 |
| PC (18:0/20:5) | M+H | 499.05 ± 424.38 | 1065.36 ± 997.00 | <0.0001 | 2.00 |
| PC (16:0/22:5) | M+H | 5017.37 ± 1245.70 | 5773.87 ± 1391.83 | 0.0001 | 2.15 |
| PC (18:0/22:6) | M+H | 5048.24 ± 1735.41 | 6283.42 ± 2299.50 | <0.0001 | 2.78 |
Abbreviations: EAD, early allograft dysfunction; lysoPC, lysophosphatidylcholines; PC, phosphatidylcholine.
Discriminative abilities of betaine, LysoPC (16:0), PC (38:5) and palmitic acid (C16:0) as predictors for EAD and short-term mortality
Betaine, palmitic acid, lysoPC (16:0) and PC (18:0/20:5) were tested for their discriminative ability as potential biomarkers for the early prediction of EAD. The area under the curve (AUC) of ROC curve of the 4 selected metabolites and five other parameters including lactate dehydrogenase (LDH), ALT, AST, bilirubin, and INR were calculated (Table 4). Individual metabolites betaine, palmitic acid, lysoPC (16:0) and PC (18:0/20:5) per se showed AUCs of 0.6855, 0.6406, 0.6806 and 0.6304, respectively. The AUC for a combination of the aforementioned metabolites increased to 0.8210 (Figure 2A), suggesting a potential of these lipidomics-derived biomarkers in predicting EAD after LDLT as early as postoperative day 1. As a panel to predict EAD, four metabolites in addition to total bilirubin further augmented AUC to 0.846. Models for other secondary outcomes were similarly constructed as shown in Figure 2B and 2C.
Table 4.
Receiver operating characteristic (ROC) curve analysis for individual metabolites in the prediction of EAD
| AUC | Standard error | |
|---|---|---|
| Betaine | 0.686 | 0.0394 |
| PC (18:0/20:5) | 0.630 | 0.0392 |
| LysoPC (16:0) | 0.681 | 0.0404 |
| Palmitic acid (C16:0) | 0.641 | 0.0401 |
| Combination of betaine, PC (18:0/20:5), palmitic acid and lysoPC (16:0) | 0.821 | 0.0336 |
| Total bilirubin | 0.754 | 0.0438 |
| Combination of metabolites and total bilirubin | 0.846 | 0.0314 |
Abbreviations: EAD, early allograft dysfunction; PC, phosphatidylcholine; lysoPC, lysophosphatidylcholine; LDH, lactate dehydrogenase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; INR, internationalized ratio.
Figure 2.
Prediction of (A) EAD, (B) the all-cause in-hospital mortality and (C) the mortality within 7 days after LDLT. (A) A combination of betaine, PC (18:0), lysoPC (16:0) and palmitic acid (C16:0) gives an AUC of 0.821 in the prediction of EAD. A panel consisting of the metabolites and total bilirubin gives an AUC of 0.846. (B) A combination of betaine, PC (18:0), lysoPC (16:0) and palmitic acid (C16:0) gives an AUC of 0.786 in the prediction of in-hospital mortality. A panel consisting of the metabolites and total bilirubin gives an AUC of 0.843. (C) A combination of betaine, PC (18:0), lysoPC (16:0) and palmitic acid (C16:0) gives an AUC of 0.958 in the prediction of 7-day mortality. A panel consisting of the metabolites and total bilirubin gives an AUC of 0.954. Abbreviations: LDLT, living donor liver transplantation; PC, phosphatidylcholine; lysoPC, lysophosphatidylcholines; AUC, the area under the receiver operator characteristic curve.
Discussion
As LDLT has become widely accepted, the early identification of poor function of liver allograft is of imminent importance for the transplant surgeons as EAD is often associated with allograft loss or mortality [22]. Previously, we have identified biomarkers that may be associated with early allograft dysfunction on postoperative day 7. In the present study, we have evaluated the metablomics using LC-MS between EAD and non-EAD recipients as early as postoperative day 1. Lipidomic analysis may provide clinicians information on the function and cellular state of a graft. The ROC analysis revealed that a combination of these plasma lipid molecules and clinically available parameters may serve as an excellent early predictor panel for EAD and early mortality, which may become applicable to clinical practice.
Betaine is a stable and natural substance existing in biological systems. The original betaine trimethylglycine, a derivative of choline, is involved in transmethylation reactions, and is mainly present in kidneys and liver. In kidneys, betaine serves as an osmolyte that balances the high extracellular osmolarity and maintains normal cell volume. In liver, via betaine-homocysteine methytransferase (BHMT), betaine transfers a methyl group to homocysteine to form methionine, which in turn can form S-adenosylmethionine (SAM), another methylating agent critical to maintenance of the integrity of the liver. A schematic illustration of metabolomic disturbances associated with liver function after LDLT is depicted in Figure 3. SAM converts phosphatidylethanolamine (PE) to phosphatidylcholine (PC), which is an integral component of lipoproteins. Betaine increases the transmethylation rate of methionine, homocysteine re-methylation and oxidation of methionine in healthy adults [23]. In humans, decreased betaine levels were associated with non-alcoholic steatohepatitis [24]. On the other hand, increased betaine levels secondary to a diminished SAM level has been observed in individuals with chronic alcohol abuse, as the hepatocytes cannot replenish SAM via the BHMT pathway [25]. Dietary betaine treatment appears to alleviate the fatty liver in such patients [26]. Impairment of BHMT pathways may increase homocysteine, which contributes to the development of liver steatosis and injury, suggesting the importance of the betaine/BHMT system in maintenance of the homeostasis of homocysteine and methionine [27]. In animal studies, bhmt-/- mice had an elevated plasma homocysteine concentration and a reduced hepatic methionine to homocysteine ratio. Deletion of bhmt gene also led to an accumulation of betaine and diminishment of choline, phosphocholine and phosphatidylcholines in tissues. Additionally, histopathological analysis of these bhmt-/- mice developed hepatocellular carcinoma or carcinoma precursors [28]. In this study, betaine level is elevated in the EAD group. Further investigation is needed to elucidate whether this betaine level is related to abnormal BHMT expression owing to recipients’ liver status.
Figure 3.
Schematic illustration of metabolomic disturbances associated with liver function after living donor liver transplantation. A decrease in PC and lysoPC species with an increase in FFA and betaine were associated with EAD and mortality. Abreviations: EAD, early allograft dysfunction; PC, phosphatidylcholine; FFA, free fatty acid; DMG, dimethylglycine; SAM, S-adenosylmethionine.
In our study, a few lipids, more specifically, palmitic acid (C16:0), lysoPC (16:0) and PC (18:0/20:5), were found to effectively differentiate EAD from non-EAD on postoperative day 1. Free fatty acids (FFA), derived from catabolism of triglyceride, play important roles in the synthesis of signalling molecules and complex lipids [29]. Serum FFA levels are elevated in obese patients. Palmitic acid (PA) is more hepato-toxic than other saturated and unsaturated FFA, and is implicated in the pathogenesis of non-alcoholic fatty liver disease and liver fibrosis [30-32]. Lysophosphatidylcholine is another species of lipids that are associated with the progression of liver disease [33] and inflammation status [34]. A relationship between the decrease in abundance of long chain lysoPC species and the progression of HBV-associated liver disease has also been established [35]. Consistent with such findings, lysoPC (16:0) and (18:0) may have a role in signalling liver tissue damage and early allograft dysfunction in patients undergoing liver transplantation [36].
Phosphatidycholine is synthesized by the two pathways, namely the CDP-choline pathway and the conversion from phosphatidylethanolamine (PE). In CDP-choline pathway, choline entering the cell is rapidly phosphorylated to phosphocholine via choline kinase, followed by the conversion of phosphocholine to CDP-choline via CTP: phosphocholine cytidylyltransferase (CT). Phosphocholine is transferred from CDP-choline to diacylglycerol (DAG) by CDP-choline: 1,2-diacylglycerol cholinephosphatransferase (CPT) to form phosphatidylcholine. This reaction occurs essentially only in hepatocytes in mammals. PCs have been proposed as a risk and prognostic biomarkers for different liver disease. For example, PC (16:0/16:0) and PC (16:0/18:0) were elevated in liver cirrhosis patients with and without HCC [36]. However, PEMT that preferentially synthesizes long chain polyunsaturated PC is downregulated in expression in HCC patients, leading to a specific decrease in such phospholipid species. Consistent with this, we have found that patients of the EAD group had higher level of palmitic acid and lower levels of lysoPC (16:0) and long chain PC than those of non-EAD group, suggesting that such changes can be used as early predictors for the development of EAD. As previously mentioned, EAD is diagnosed on postoperative day 7 based on live function and coagulation profiles. We have demonstrated that, as early as postoperative day 1, both clinical parameters and metabolomic biomarkers may be adequate in predicting EAD on postoperative day 7; however, a panel consisting of metabolites and bilirubin performs more superior than either alone. Limitations still apply to the study. The small number of patients can be the potential limitation and validation in a larger data set is warranted. Additional research is also required to pursue the mechanisms of these identified biomarkers in association with the dysfunction and mortality after LDLT.
Conclusions
The combination of abovementioned four metabolites in patients on postoperative day 1 is highly predictive of EAD with AUCs of 0.8210, better than the other clinically available markers. The predictive power of metabolites is further improved when total bilirubin is added to the calculation as a panel. In addition, this panel exhibits a high prognostic accuracy in the prediction of all-cause in-hospital mortality and mortality within 7 postoperative days with AUCs of 0.843 and 0.954. The panel consisting of betaine, palmitic acid, PC and lysoPC species and total bilirubin demonstrates the power in in diagnostic and prognostic evaluation of LDLT.
Acknowledgements
This work was supported by Chang Gung Memorial Hospital (CMRPG3H1191-3).
Disclosure of conflict of interest
None.
References
- 1.Goldaracena N, Gorgen A, Doyle A, Hansen BE, Tomiyama K, Zhang W, Ghanekar A, Lilly L, Cattral M, Galvin Z, Selzner M, Bhat M, Selzner N, McGilvray I, Greig PD, Grant DR, Sapisochin G. Live donor liver transplantation for patients with hepatocellular carcinoma offers increased survival vs. deceased donation. J Hepatol. 2019;70:666–673. doi: 10.1016/j.jhep.2018.12.029. [DOI] [PubMed] [Google Scholar]
- 2.Wong TCL, Ng KKC, Fung JYY, Chan AAC, Cheung TT, Chok KSH, Dai JWC, Lo CM. Long-term survival outcome between living donor and deceased donor liver transplant for hepatocellular carcinoma: Intention-to-treat and propensity score matching analyses. Ann Surg Oncol. 2019;26:1454–1462. doi: 10.1245/s10434-019-07206-0. [DOI] [PubMed] [Google Scholar]
- 3.Neves DB, Rusi MB, Diaz LG, Salvalaggio P. Primary graft dysfunction of the liver: definitions, diagnostic criteria and risk factors. Einstein (Sao Paulo) 2016;14:567–572. doi: 10.1590/S1679-45082016RW3585. [DOI] [PubMed] [Google Scholar]
- 4.Chen XB, Xu MQ. Primary graft dysfunction after liver transplantation. Hepatobiliary Pancreat Dis Int. 2014;13:125–137. doi: 10.1016/s1499-3872(14)60023-0. [DOI] [PubMed] [Google Scholar]
- 5.Olthoff KM, Kulik L, Samstein B, Kaminski M, Abecassis M, Emond J, Shaked A, Christie JD. Validation of a current definition of early allograft dysfunction in liver transplant recipients and analysis of risk factors. Liver Transpl. 2010;16:943–949. doi: 10.1002/lt.22091. [DOI] [PubMed] [Google Scholar]
- 6.Deschenes M. Early allograft dysfunction: causes, recognition, and management. Liver Transplant. 2013;19:S6–S8. doi: 10.1002/lt.23746. [DOI] [PubMed] [Google Scholar]
- 7.Alvarez-Mercado AI, Gulfo J, Gomez MR, Jimenez-Castro MB, Gracia-Sancho J, Peralta C. Use of steatotic grafts in liver transplantation: current status. Liver Transpl. 2019;25:771–786. doi: 10.1002/lt.25430. [DOI] [PubMed] [Google Scholar]
- 8.Olthoff KM, Smith AR, Abecassis M, Baker T, Emond JC, Berg CL, Beil CA, Burton JR Jr, Fisher RA, Freise CE, Gillespie BW, Grant DR, Humar A, Kam I, Merion RM, Pomfret EA, Samstein B, Shaked A. Defining long-term outcomes with living donor liver transplantation in North America. Ann Surg. 2015;262:465–475. doi: 10.1097/SLA.0000000000001383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tomescuand D, Popescu M, Dima SO. Rotational thromboelastometry (ROTEM) 24 hours post liver transplantation predicts early allograft dysfunction. Rom J Anaesth Intensive Care. 2018;25:117–122. doi: 10.21454/rjaic.7518.252.tms. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Haugen CE, Holscher CM, Luo X, Bowring MG, Orandi BJ, Thomas AG, Garonzik-Wang J, Massie AB, Philosophe B, McAdams-DeMarco M, Segev DL. Assessment of trends in transplantation of liver grafts from older donors and outcomes in recipients of liver grafts from older donors, 2003-2016. JAMA Surg. 2019;154:441–449. doi: 10.1001/jamasurg.2018.5568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Durand F, Levitsky J, Cauchy F, Gilgenkrantz H, Soubrane O, Francoz C. Age and liver transplantation. J Hepatol. 2019;70:745–758. doi: 10.1016/j.jhep.2018.12.009. [DOI] [PubMed] [Google Scholar]
- 12.Safaei A, Arefi Oskouie A, Mohebbi SR, Rezaei-Tavirani M, Mahboubi M, Peyvandi M, Okhovatian F, Zamanian-Azodi M. Metabolomic analysis of human cirrhosis, hepatocellular carcinoma, non-alcoholic fatty liver disease and non-alcoholic steatohepatitis diseases. Gastroenterol Hepatol Bed Bench. 2016;9:158–173. [PMC free article] [PubMed] [Google Scholar]
- 13.Tsai HI, Lo CJ, Zheng CW, Lee CW, Lee WC, Lin JR, Shiao MS, Cheng ML, Yu HP. A lipidomics study reveals lipid signatures associated with early allograft dysfunction in living donor liver transplantation. J Clin Med. 2018;8:30. doi: 10.3390/jcm8010030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Chan KM, Eldeen FZ, Lee CF, Wu TJ, Chou HS, Wu TH, Soong RS, Lee WC. “Left at right” adult liver transplantation: the feasibility of heterotopic implantation of left liver graft. Am J Transplant. 2012;12:1511–1518. doi: 10.1111/j.1600-6143.2012.03997.x. [DOI] [PubMed] [Google Scholar]
- 15.Croome KP, Wall W, Quan D, Vangala S, McAlister V, Marotta P, Hernandez-Alejandro R. Evaluation of the updated definition of early allograft dysfunction in donation after brain death and donation after cardiac death liver allografts. Hepatobiliary Pancreat Dis Int. 2012;11:372–376. doi: 10.1016/s1499-3872(12)60194-5. [DOI] [PubMed] [Google Scholar]
- 16.Sarafian MH, Gaudin M, Lewis MR, Martin FP, Holmes E, Nicholson JK, Dumas MC. Objective set of criteria for optimization of sample preparation procedures for ultra-high throughput untargeted blood plasma lipid profiling by ultra performance liquid chromatography-mass spectrometry. Anal Chem. 2014;86:5766–5774. doi: 10.1021/ac500317c. [DOI] [PubMed] [Google Scholar]
- 17.Jacquenod P, Wallon G, Gazon M, Darnis B, Pradat P, Virlogeux V, Farges O, Aubrun F. Incidence and risk factors of coagulation profile derangement after liver surgery: implications for the use of epidural analgesia-A retrospective cohort study. Anesth Analg. 2018;126:1142–1147. doi: 10.1213/ANE.0000000000002457. [DOI] [PubMed] [Google Scholar]
- 18.Lai Q, Giovanardi F, Melandro F, Larghi Laureiro Z, Merli M, Lattanzi B, Hassan R, Rossi M, Mennini G. Donor-to-recipient gender match in liver transplantation: a systematic review and meta-analysis. World J Gastroenterol. 2018;24:2203–2210. doi: 10.3748/wjg.v24.i20.2203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lee EC, Kim SH, Park SJ. Outcomes after liver transplantation in accordance with ABO compatibility: a systematic review and meta-analysis. World J Gastroenterol. 2017;23:6516–6533. doi: 10.3748/wjg.v23.i35.6516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Dahm F, Georgiev P, Clavien PA. Small-for-size syndrome after partial liver transplantation: definition, mechanisms of disease and clinical implications. Am J Transplant. 2005;5:2605–2610. doi: 10.1111/j.1600-6143.2005.01081.x. [DOI] [PubMed] [Google Scholar]
- 21.Goldaracena N, Echeverri J, Selzner M. Small-for-size syndrome in live donor liver transplantation-pathways of injury and therapeutic strategies. Clin Transplant. 2017;31:e12885. doi: 10.1111/ctr.12885. [DOI] [PubMed] [Google Scholar]
- 22.Lee DD, Croome KP, Shalev JA, Musto KR, Sharma M, Keaveny AP, Taner CB. Early allograft dysfunction after liver transplantation: an intermediate outcome measure for targeted improvements. Ann Hepatol. 2016;15:53–60. doi: 10.5604/16652681.1184212. [DOI] [PubMed] [Google Scholar]
- 23.Storch KJ, Wagner DA, Young VR. Methionine kinetics in adult men: effects of dietary betaine on L-[2H3-methyl-1-13C]methionine. Am J Clin Nutr. 1991;54:386–394. doi: 10.1093/ajcn/54.2.386. [DOI] [PubMed] [Google Scholar]
- 24.Sookoian S, Puri P, Castaño GO, Scian R, Mirshahi F, Sanyal AJ, Pirola CJ. Nonalcoholic steatohepatitis is associated with a state of betaine-insufficiency. Liver Int. 2017;37:611–619. doi: 10.1111/liv.13249. [DOI] [PubMed] [Google Scholar]
- 25.Ascha M, Wang Z, Ascha MS, Dweik R, Zein NN, Grove D, Brown JM, Maeshall S, Lopez R, Hanouneh IA. Metabolomics studies identify novel diagnostic and prognostic indicators in patients with alcoholic hepatitis. World J Hepatol. 2016;8:499–508. doi: 10.4254/wjh.v8.i10.499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Yang W, Huang L, Gao J, Wen S, Tai Y, Chen M, Huang Z, Liu R, Tang C, Li J. Betaine attenuates chronic alcoholinduced fatty liver by broadly regulating hepatic lipid metabolism. Mol Med Rep. 2017;16:5225–5234. doi: 10.3892/mmr.2017.7295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Finkelstein JD. Methionine metabolism in liver diseases. Oxford University Press; 2003. [DOI] [PubMed] [Google Scholar]
- 28.Teng YW, Mehedint MG, Garrow TA, Zeisel SH. Deletion of betaine-homocysteine S-methyltransferase in mice perturbs choline and 1-carbon metabolism, resulting in fatty liver and hepatocellular carcinomas. J Biol Chem. 2011;286:36258–36267. doi: 10.1074/jbc.M111.265348. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Di Pasquale MG. The essentials of essential fatty acids. J Diet Suppl. 2009;6:143–161. doi: 10.1080/19390210902861841. [DOI] [PubMed] [Google Scholar]
- 30.Ricchi M, Odoardi MR, Carulli L, Anzivino C, Ballestri S, Pinetti A, Fantoni LI, Marra F, Bertolotti M, Banni S, Lonardo A, Carulli N, Loria P. Differential effect of oleic and palmitic acid on lipid accumulation and apoptosis in cultured hepatocytes. J Gastroenterol Hepatol. 2009;24:830–840. doi: 10.1111/j.1440-1746.2008.05733.x. [DOI] [PubMed] [Google Scholar]
- 31.Cansancao K, Silva Monteiro L, Carvalho Leite N, Davalos A, Tavares do Carmo MDG, Peres WAF. Advanced liver fibrosis is independently associated with palmitic acid and insulin levels in patients with non-alcoholic fatty liver disease. Nutrients. 2018;10:1586. doi: 10.3390/nu10111586. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ogawa Y, Imajo K, Honda Y, Kessoku T, Tomeno W, Kato S, Fujita K, Yoneda M, Saito S, Saigusa Y, Hyogo H, Sumida Y, Itoh Y, Eguchi K, Yamanka T, Wada K, Nakajima A. Palmitate-induced lipotoxicity is crucial for the pathogenesis of nonalcoholic fatty liver disease in cooperation with gut-derived endotoxin. Sci Rep. 2018;8:11365. doi: 10.1038/s41598-018-29735-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Karen Cotte A, Cottet V, Aires V, Mouillot T, Rizk M, Vinault S, Binquet C, de Barros JPP, Hillon P, Delmas D. Phospholipid profiles and hepatocellular carcinoma risk and prognosis in cirrhotic patients. Oncotarget. 2019;10:2161–2172. doi: 10.18632/oncotarget.26738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Taylor LA, Arends J, Hodina AK, Unger C, Massing U. Plasma lyso-phosphatidylcholine concentration is decreased in cancer patients with weight loss and activated inflammatory status. Lipids Health Dis. 2007;6:17. doi: 10.1186/1476-511X-6-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wu T, Zheng X, Yang M, Zhao A, Li M, Chen T, Panee J, Jia W, Ji G. Serum lipid alterations identified in chronic hepatitis B, hepatitis B virus-associated cirrhosis and carcinoma patients. Sci Rep. 2017;7:42710. doi: 10.1038/srep42710. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Xu J, Casas-Ferreira AM, Ma Y, Sen A, Kim M, Proitsi P, Shkodra M, Tena M, Srinivasan P, Heaton N, Jassem W, Legido-Quigley C. Lipidomics comparing DCD and DBD liver allografts uncovers lysophospholipids elevated in recipients undergoing early allograft dysfunction. Sci Rep. 2015;5:17737. doi: 10.1038/srep17737. [DOI] [PMC free article] [PubMed] [Google Scholar]