Each patient’s background is summarized in Table 1. The average age of LDLT patients was 48±13 years, DDLT was 45±12, and SLKT was 46±5. The Model for End-stage Liver Disease (MELD) score was 18±6 (LDLT), 18.6±5 (DDLT), with SLKT patients having the highest score (19±3). Coagulability on the day before surgery tended to be low in DDLT patients, mild in LDLT patients, and almost standard value in SLKT patients.

Preoperative eGFR was relatively well-maintained in LDLT patients (71±39), while it showed a slight decrease in DDLT patients (52±22). The percentages of patients undergoing preoperative dialysis were 11% in LDLT and 14% in DDLT, and dialysis was performed in all SLKT patients. Of the seven LDLT patients on preoperative dialysis, five patients had chronic renal failure and two patients had an acute renal failure due to hepatorenal syndrome. The causes of liver disorders are shown in Table 2.

Primary Biliary Cholangitis (PBC) was the most common in LDLT patients (19%), alcoholic and drug-induced injuries were most common in DDLT patients (28% each), and Autosomal Dominant Polycystic Kidney Disease (ADPKD) was the most common primary disease in SLKT patients. In many of these ADPKD patients, liver function was maintained, but infection of the hepatic cysts was uncontrollable.

Intraoperatively, we monitored the electrocardiogram, non-invasive blood pressure, invasive arterial blood pressure (FloTrac^TM sensor), central venous pressure (Presep, Edwards Lifescience, USA), pulse oximeter, exhaled gas concentration, and body temperature (through bladder catheter). We used the Bispectral Index monitor (BIS monitor, Vista A-3000^TM, NIHON KODEN, Tokyo, Japan) or the Patient State Index (PSI, Sedline^®, Masimo, USA) for electroencephalogram reading. We used Transesophageal Echocardiography (TEE) in 22% of patients and measured pulmonary arterial pressure in 3% of patients.

The intraoperative anesthetic management is summarized in Table 3. The operative time was 682±114 minutes for LDLT, 690±113 minutes for DDLT, while the time for SLKT was longer by 80 minutes because renal transplantation was also required.

All the patients received general anesthesia. Two LDLT patients and one SLKT patient were intubated preoperatively in the ICU, while the other patients received propofol or midazolam, fentanyl, and remifentanil during induction. All patients were anesthetized using inhalation anesthetic agents after intubation (Table 3). We checked the anesthetic depth of the inhalation agents with BIS or PSI and adjusted it at 0.5–1 MAC (Minimum Alveolar Concentration). We controlled the intraoperative BIS and PSI values at 25–60 and 20–40, respectively.

We used inhalation anesthetic agents sevoflurane and desflurane (64% vs. 35% in LDLT; 14% vs. 85% in DDLT). We administered propofol for induction in 82% of patients in small divided doses of 0.8–2 mg/kg according to blood pressure and midazolam (0.05–0.2 mg/kg) in 17% of patients. We used rocuronium for muscle relaxation in all patients, which we continuously administered in 68% of LDLT patients. We monitored the infusion rate at a median of 6 (3–8) μg/kg/min with TOF-Watch^® and changed it as needed. Since we were using the train-of-four count, we recorded a few points for each patient but could not evaluate muscle relaxation.

Since there was a lot of bleeding and ascites, the water balance was relatively low (+2.6–4.5 mL/kg/h) in all patients but was relatively high in SLKT patients (5.5 mL/kg/h). The urine volume was maintained at 1.5 mL/kg/h in LDLT, and the rate of diuretic use (human atrial natriuretic peptide [hANP], furosemide) was high. The hANP was continuously administered during induction without dialysis in patients with a preoperative eGFR of 20–40. Furosemide was used as a single dose at the end of surgery in patients with a preoperative eGFR of 60–100. For postoperative management to facilitate renal function, we increased the infusion until diuresis while draining pleural effusion and ascites.

During infusion, the albumin dosage was 32% of the total infusion and transfusion volume (Fig. 1). During a blood transfusion, Red Blood Cells (RBCs), Fresh Frozen Plasma (FFP), and platelets were used in larger quantities in DDLT than in LDLT patients; however, the doses of FFP and platelets were adjusted depending on individual patients.

Fig. (1). Proportion of total infusion and transfusion volumes (LDLT) (n=51). *RBC: red blood cell, **FFP: frozen fresh plasma, ***PC: platelet concentrates, Every item is the median of all the LDLT cases.

Catecholamine was used to maintain the mean Arterial Blood Pressure (ABP) at 40–80 mmHg. Adrenaline was used in a few patients (11% in LDLT, 14% in DDLT, 20% in SLKT), but dopamine and noradrenaline were used in all patients. On comparing the median values, the doses of dopamine and noradrenaline were found to be higher in SLKT patients (Table 3).

The value of graft-to-recipient weight ratio varied from 0.54 to 2.36 (median = 0.97).

The patients’ postoperative courses are shown in Table 4. The hospitalization period ascendingly increased in LDLT, DDLT, and SLKT patients, but the length of ICU stay was shortest in DDLT patients. Of 63, six patients died, and all underwent LDLT. Of these, four patients died within one year.

Postoperative dialysis was introduced in two LDLT patients, but it was not directly related to intraoperative management. Dialysis was withdrawn in four of five SLKT patients and all patients with preoperative hepatorenal syndrome. On the day after surgery, eGFR was maintained at the same value as what was compared with the rate before surgery in all transplantation methods, and it further improved at the time of discharge. The infection rate was 25–28% in LDLT and DDLT patients, whereas 0% in SLKT patients. The rejection rate was 23% in LDLT and 14% in DDLT patients, 0% in SLKT patients.

The patients were extubated at an average of a day after transplantation. There was no induction treatment with antibodies. The basic triple-drug regimen of immunosuppression was followed: tacrolimus, steroid, and Mycophenolate Mofetil (MMF). The trough level of tacrolimus was 7–9 ng/dl during the first month. Almost all patients with LDLT showed massive ascites. In order to avoid dehydration, the entire volume of drained ascites was meticulously compensated with crystalloid and 5% albumin. The compensated volume was decreased until the amount of ascites decreased and graft growth was seen. It took 3-4 weeks for graft growth and the patients stayed in the ICU and HCU during this period of care.

The average patient age was 40 years. In recent times, the age of LT recipients in the USA has gradually increased; now, patients aged 50–64 years account for nearly half of all recipients (UNOS (United Network for Organ Sharing), https://optn.transplant.hrsa.gov/). The mean age is expected to increase if recipients with alcoholic cirrhosis and fatty liver increase in the future. LT for recipients with heart disease and diabetes mellitus will require further technical improvements and data accumulation.

In this study, the MELD score was approximately 18–19 (lowest=7; highest=34). Many SLKT patients had uncontrollable hepatic cyst infections due to ADPKD. Therefore, although the hepatic function was preserved, the MELD score in these patients remained high because of renal failure.

All patients received general anesthesia, and the postoperative pain was controlled using fentanyl via intravenous patient-controlled analgesia (iv-PCA). Epidural anesthesia was not performed to avoid coagulation abnormality. Volatile anesthetic agents are hardly metabolized in the body, and their preconditioning effect has been reported to be organ-protective [2]. In this study, 38 patients received sevoflurane (60%), and 25 received desflurane (40%). In a randomized control trial comparing the incidence of Postreperfusion Syndrome (PRS) between sevoflurane and desflurane in LDLT, Lee et al. reported that PRS was significantly higher in the desflurane group (38.7 vs. 77.4, p = 0.004). Further, the desflurane group had heavier bleeding and used more epinephrine [3]. However, in another study, postoperative hepatorenal function and coagulability in LT donors were better with desflurane than with sevoflurane [4]. At our hospital, the anesthesiologist-in-charge determines the use of the inhalation agent according to each case, and thus, no clear criteria have been established. Since desflurane needs to be frequently filled in the vaporizer, some anesthesiologists avoid it.

Rocuronium was used as a muscle relaxant in all patients and was continuously administered to nearly 70% of patients. Many anesthesiologists perform continuous administration because they cannot afford to administer single doses frequently. Since rocuronium is metabolized in the liver, the continuous administration rate was increased or decreased using a muscle relaxation monitor. The dose used in our study was close to the usual dose (3–8 μg/kg/min) to avoid excessive bleeding that could result in unstable plasma drug concentration.

A pulmonary artery catheter was inserted in 2 patients (3%) with pulmonary hypertension. In other cases, cardiac output was estimated and managed only with ABP and FloTrac. TEE was useful not only for cardiac function assessment but also for infusion/transfusion volume load management. The American Society of Anesthesiologists also recommends TEE during LT. However, patients with liver failure often have esophageal varices, and TEE cannot be performed in such patients. In this study, TEE was used in 22% of patients. In one patient, it was difficult to extubate on the day after surgery due to lingual hematoma caused by TEE complications.

The BIS and Sedline were used as EEG monitors during anesthesia. Although there have been many reports about Postoperative Cognitive Dysfunction (POCD) with BIS-guided anesthesia [5-7], few studies have been conducted in patients with LT [8]. In this study, the delivery of anesthetic agents was managed by dose adjustment based on the intraoperative BIS/PSI values in all patients. POCD was found in 1 of 63 patients; however, it was described subjectively only in the medical record, and no delirium score was used. In many cases, rSO₂ could not be measured due to jaundice associated with hyperbilirubinemia.

In end-stage liver failure, the Renin-Angiotensin-Aldosterone System (RAAS) is enhanced and is likely to cause the hepatorenal syndrome. In this study, all patients undergoing preoperative dialysis due to acute kidney injury were successfully withdrawn from dialysis after surgery, as appropriate transfusion management and the use of hANP were expected to restore renal function.

Regarding intraoperative diuretic use, hANP was continuously administered to patients with decreased renal function and self-urination, and furosemide was used in patients with decreased urine volume despite the maintenance of renal function. hANP promotes natriuresis by dilating afferent arterioles in the kidneys, increasing renal blood flow, and suppressing RAAS hormones. RAAS is enhanced in patients with terminal cirrhosis due to decreased renal blood flow; the use of hANP antagonizes this condition.

Patients with terminal cirrhosis develop ascites and pleural effusion due to hypoalbuminemia and decreased effective circulating blood volume. Peripheral vascular resistance decreases due to dysfunction of endothelin and nitrous oxide and is further reduced by general anesthetic agents, which increases the risk of circulatory collapse. Additionally, coagulation disorders cause intraoperative blood loss, and hypotension is likely to occur during surgery after anesthesia induction. The use of catecholamines is essential to combat this.

Hypotension during LT can affect the induction of anesthesia, drainage of ascites, clumping of the Inferior Vena Cava (IVC), and reperfusion. It is important to communicate with surgeons at each time point and control the dosage of catecholamines. In many of our patients, catecholamines (noradrenaline and dopamine) were continuously administered after central venous access was secured. In 3 patients with less blood loss, the condition was controlled only with single doses of ephedrine and phenylephrine. However, it is difficult to predict intraoperative hemodynamics in LT, and it was considered reasonable to start continuous catecholamine administration in small amounts to correspond to hypotension immediately, even in patients with no hypotension at the time of anesthesia induction. Especially during the IVC-clumping period, catecholamines (0.05 - 0.3 mcg/kg/min of noradrenaline and 3 - 5 mcg/kg/min of dopamine) were mainly used for maintaining blood pressure rather than just increasing infusion. Their dosage was based on cardiac output, cardiac index, and a mean ABP of at least 40 mmHg. Avoiding perfusion of excess fluid could protect the donor's liver and reduce respiratory complications. In some DDLT patients, adrenaline was required because abnormal hypotension due to reperfusion injury may have occurred if the duration of ischemia in the donor's liver was prolonged. No complications caused by intraoperative hypotension were observed.

The colloid infusions, 5% albumin and HES, were used (except for one patient in whom albumin was not used). If the colloid osmotic pressure did not improve after surgery, the intubation period was prolonged to avoid pulmonary edema or pleural effusion. Enough amount of albumin was used to significantly improve the patients’ lung condition. The albumin infusion volume was over 30% of the total infusion and transfusion volume (Fig. 1). Although albumin infusions are expensive, they might be reasonable for maintaining intravascular volume, osmotic pressure, and protecting renal function. For RBC transfusion, a necessary volume was transfused while checking blood pressure, Hemoglobin (Hb), and Hematocrit (Ht). Since high Ht increases the risk of thrombosis, RBC was administered to maintain a volume of around 30%. Further, the hemostatic coagulation system gets disrupted in end-stage liver disease. Therefore, bleeding and embolism were likely to occur. It has been reported that transfusion of FFP and platelets are more likely to cause portal vein thrombosis [9]. Therefore, in our hospital, we try not to transfuse FFP or platelets as far as possible, as was the case in this study.