Abstract

Objectives

Liver retransplantation (Re-LT) is the effective therapy for irreversible liver graft failure after primary liver transplantation (LT). The challenges faced by the operative team in the Re-LT setting have been seldom elucidated. Our aim is to analyze the differences in fluid management in primary LT and Re-LT during the surgical procedure.

Methods

The anesthesia charts of 16 patients who underwent both primary LT and Re-LT at our center in the space from October 1995 to May 2009 were analyzed. Group 1 (GI) consisted of patients who underwent primary LT, whereas patients in Group 2 (GII) were patients in GI but underwent Re-LT. GI was further divided into two subgroups depending on whether they had previous abdominal surgery before primary LT (GIB) or not (GIA). Wilcoxon signed-ranks test was used to compare GI and GII, and GIA and GIB. A p value less than 0.05 was regarded as significant. Data were given as mean ± standard deviation.

Results

Blood loss was significantly increased from 48.9 ± 106 mL/kg in GI to 251.5 ± 242 mL/kg in GII. Consequently more blood products, crystalloids, sodium bicarbonate, calcium chloride, and neosynephrine were required to support the hemodynamics in GII. In GI, GIB tended to bleed more and required more blood transfusions than GIA.

Conclusion

More bleeding is expected in Re-LT than primary LT. Additional anesthetic personnel, more intravenous lines, and blood and blood products should be readily available to deal with the emergent fluid and hemodynamic resuscitations in anesthesia for Re-LT.

Keywords

anesthesia, general; blood loss, surgical; fluid therapy; liver transplantation;

1. Introduction

Liver transplantation (LT) is the standard surgical therapy for terminal liver failure.¹ One and 5-year survival rates as reported are higher than 80% and higher than 70%, respectively, in experienced centers.2, 3, 4, 5 In biliary atresia patients, the 5-year survival rate after LT was 98%.⁵ However, despite successful LT, the liver graft may be subject to complications, damage from operative insults, or disease recurrence, which may lead to irreversible graft dysfunction that necessitate liver retransplantation (Re-LT).⁶ The outcome of Re-LT as reported is significantly inferior to that of primary LT.7, 8, 9 This may implicate that Re-LT is more challenging for the transplant team. Anesthesia management in LT is already a challenge to most anesthesiologists but whether Re-LT is a much bigger challenge remains to be elucidated. There is still limited information about the comparison of the anesthesia management in LT and Re-LT in the same patients. The purpose of this study is to analyze the differences in fluid management in primary LT and Re-LT during the surgical procedure.

2. Methods

Between October 1995 and May 2009, there were 16 patients who underwent both primary LT and Re-LT at Kaohsiung Chang Gung Memorial Hospital, Taiwan. The study protocol was approved by the Institutional Review Board for Human Studies of Chang Gung Memorial Hospital. The anesthesia records were analyzed retrospectively. Preoperative hemoglobin (Hb), coagulation profile, renal function, and albumin were recorded. All patients were monitored on electrocardiography, continuous arterial blood pressure, and pulse oximetry. Central venous pressure, end-tidal CO₂, body temperature, and urine output were, likewise, monitored. General anesthesia was induced with propofol and fentanyl and maintained with isoflurane in an O₂/air mixture combined with fentanyl when required.

Atracurium or cisatracurium was used as muscle relaxant. The LT operation was performed pursuant to standard surgical technique without venovenous bypass. Dopamine 2 μg/kg/hr was given continuously throughout the operation. Blood recirculating apparatus was not used and antifibrinolytics were not given during operation. During the LT procedure, based on the findings and recommendations by Massicotte et al,¹⁰ coagulation profile monitoring was not routinely performed. Arterial blood gas analysis was done at least five times throughout the procedure (after induction of anesthesia, 2^nd and/or 4^th hour through the operation, during anhepatic phase, 10 minutes after portal reperfusion, and at the end of the operation). Metabolic acidosis was corrected with 7% sodium bicarbonate when the base excess was greater than −5%, and 5% calcium chloride was administered to treat ionized hypocalcemia when the serum ionized calcium was lower than 0.8 mmol/L.¹¹ Crystalloid and 5% albumin as colloid, rather than blood products, were used as blood loss replacement to maintain normovolemia. Blood transfusion was started when the Hb level was lower than 6–7 g/dL. When blood transfusion was required, leukocyte-poor red blood cells were given in amounts calculated to give a post-transfusion target Hb of 9 g/dL. Fresh frozen plasma and platelets were transfused only when blood clotting was not appreciated in the surgical field. The total amount of crystalloid solutions, 5% albumin, leukocyte-poor red blood cells, fresh frozen plasma, and platelets given were recorded, analyzed, and compared.

The patients were divided into two groups: Group 1 (GI, Primary LT) and Group 2 (GII, Re-LT). GI patients were further divided into two subgroups depending on whether they had previous abdominal surgery before primary LT (GIB) or not (GIA).

Because both pediatric and adult patients were included in this series, the data on amount of blood loss, fluid replacements, and blood and blood product transfusions given were converted into milliliter/kilogram for statistical comparison. All data were presented as mean ± standard deviation. Because the continuous data could not fit the assumption of normal distribution, Wilcoxon signed-ranks test was used. The difference was considered significant when the p value was less than 0.05.

3. Results

Sixteen patients underwent primary LT (3 deceased donor LT, 13 living donor LT). There were nine male and seven female patients. Table 1 shows the characteristics of the 16 patients. Eight patients had abdominal surgery before primary LT, which included Kasai operation, hepatectomy, cholecystectomy, and cesarean section. The indications for Re-LT are shown in Table 1. Table 2 shows the demographical data, baseline laboratory values, and intraoperative parameters of the 16 patients at primary LT (GI) and at re-LT (GII). There were no significant differences found in the patients’ preoperative data.

Table 1. Characteristics of 16 patients.

Download full-size image

Table 2. Comparison of demographic data, baseline laboratory values, and intraoperative parameters between primary liver transplantation (Group I) and liver retransplantation (Group II).

Download full-size image

There was no intraoperative mortality in this series. One patient developed intraoperative cardiac arrest that required cardiopulmonary resuscitation. In the very same patients who underwent primary LT and then Re-LT, blood loss was significantly increased. Blood loss in Re-LT was almost five times that of primary LT (48.9 ± 106 mL/kg in primary LT vs. 251.5 ± 242 mL/kg in Re-LT) (Table 2). All in all, the blood loss in GI was less than 50% of the blood volume, whereas in the 13 of 16 patients in GII, the blood loss was more than one blood volume.

The mean volume of crystalloids used in GI and GII was 175.1 ± 110 mL/kg and 321.7 ± 175 mL/kg, respectively. The mean volume of crystalloid used was greater in GII and was statistically significant (Table 3). The blood and blood products (i.e. leukocyte-poor packed red cells and platelets) were also significantly greater in GII than GI. There was no significant difference in the Hb levels at the end of the operation between GI and GII. GII patients required more sodium bicarbonate to buffer intraoperative metabolic acidosis and more calcium chloride to treat ionized hypocalcemia. In GII, seven patients required neosynephrine and another two patients required epinephrine to support their hemodynamics. In contrast, only one patient required neosynephrine and epinephrine was not used in any patient in GI (Table 3).

Table 3. Blood loss and intraoperative transfusion requirements in the 16 primary liver transplantation (Group I) and liver retransplantation (Group II) recipients.

Download full-size image

The blood loss in GIA and GIB was 26.5 ± 24.2 mL/kg and 71.2 ± 150.3 mL/kg, respectively, the blood and fluids replacement were tended to be more in GIB (Table 4).

Table 4. Blood loss and intra-operative transfusion requirements in GIA and GIB patients.

Download full-size image

4. Discussion

Our results show that the circumstance of anesthesia management centered on blood loss and fluid replacement of the same patients undergoing primary LT and Re-LT was significantly different. The main cause of difference was in the blood loss. The blood loss was significantly more in GII than in GI. The mean blood loss of GII was almost five times more than the mean blood loss in GI (251.5 ± 242 mL/kg GII vs. 48.9 ± 106 mL/kg GI). No patient in GI had an episode of massive bleeding but 13 of the 16 patients in GII had blood loss greater than their blood volume. It also indicated that the hemodynamics of GII patients were more unstable than GI and required more concentrated monitoring. From the results, the requirements in fluid replacement, neosynephrine, sodium bicarbonate, and calcium chloride for maintaining hemodynamics were also more in GII than in GI. Despite the relatively unstable hemodynamics encountered in GII, sufficient urine output could still be generated and maintained (Table 3).

In major abdominal surgery, blood loss should be appropriately replaced. The initial bleeding is usually replaced by crystalloids in three- to fourfold the volume of blood lost. When larger amounts of bleeding are encountered or more than 20% of the total blood volume is lost, blood product(s) transfusion is indicated.⁷ Because red blood cell and platelet transfusions have negative impact on survival after LT,12, 13 we consider giving blood transfusion only when the Hb level is lower than 6–7 g/dL. Normovolemic anemia is allowed to occur in the recipient intraoperatively by allowing the Hb to decrease naturally through surgical bleeding to as low as 6–7 g/dL and, then, maintain it at about 9 g/dL postoperatively.¹⁴ This level of Hb seemed to have no negative effect on the function of vital organs, such as the brain, heart, lungs, and kidneys, and the recovery of the liver graft.¹⁴ Table 3 shows that the Hb levels at the end of the operation between groups were similar at around 9 g/dL indicating that both groups were not overtransfused.

Massive blood loss followed by massive blood transfusion and fluid resuscitation is associated with increased mortality and morbidity.¹⁵ It is even reported to be a predictor of poor prognosis in living donor LT recipients.¹⁶ Massive bleeding had a negative impact in patient survival after Re-LT in our series. One patient had cardiac arrest during the procedure and the conditions of 5 of 16 patients in GII were critical during the postoperative period and all 5 patients died. The outcome of our Re-LT was inferior to that of our previous primary LT in biliary atresia patients.⁵ Likewise, significant higher morbidity and mortality in Re-LT setting in comparison to primary LT were also reported by other centers.7, 8, 9

In the LT setting, blood loss may vary from minimal to massive.16, 17 In our pediatric LT, 42% did not require intraoperative blood transfusion.¹⁴ Similarly, Massicotte et al¹⁰ reported that 79% of their patients did not receive blood transfusion during LT. But LT associated with massive blood loss are more often reported.16, 18 Analyses show that it is difficult to predict blood loss before an LT operation.10, 19 Neither preoperative nor intraoperative clotting parameters were predictor for the intraoperative blood loss and blood products used.10, 19 Hence, what is the primary reason why Re-LT was associated with significantly more blood loss than primary LT as shown by our results? Table 2 shows that both GI and GII had prolonged international normalized ratio, activated partial prothrombin time, and thrombocytopenia but only GII had massive bleeding. These findings corroborate that preoperative coagulation profile is not a predictor and does not correlate with intraoperative blood loss.10, 19 It is, however, possible that technical factors rather than patient-related factors are more important in the control of intraoperative bleeding in the Re-LT setting.²⁰ Because the primary LT and Re-LT operations were performed by the same surgical and anesthesia teams where the only difference between the groups is the previous LT, a history of primary LT can, therefore, be deemed as a contributing factor that caused five times more bleeding in GII.

The issue of whether previous abdominal surgery affects the intraoperative blood loss or not is not settled. Ozier et al²¹ found that previous abdominal surgery was not a factor correlated with blood loss, but other authors noted its relation to blood loss.22, 23, 24 The increase in blood loss is often attributed to surgical technical problems associated with the presence of adhesions and collateral vessels after previous abdominal surgery.22, 25 Eight patients in GI had previous abdominal surgery (Table 1). When GI was divided into two subgroups, analysis showed that patients with previous abdominal surgery in the subgroup GIB tended to bleed more (71.2 ± 150.3 mL/kg in GIB vs. 26.5 ± 24.2 mL/kg in GIA) and required relatively more red blood cell (44.8 ± 83.5 mL/kg in GIB vs. 14.4 ± 9.2 mL/kg in GIA) transfusion. It is expected that the surgical technique to deal with problems related to adhesions in the inferior vena cava (IVC) and hepatic veins from previous LT are more challenging than adhesions after Kasai operation or other abdominal surgeries not involving the IVC. Our results showed that blood loss in patients without previous abdominal surgery (GIA), patients with previous non-LT abdominal surgery (GIB), and patients after LT surgery (GII) were in increasing order with blood loss at 26.5 ± 24.2 mL/kg, 71.2 ± 150.3 mL/kg, and 251.5 ± 242.5 mg/kg, respectively. The blood product transfusion requirements were consequently in increasing order at 14.4 ± 9.2 mL/kg, 44.8 ± 83.5 mL/kg, and 151.8 ± 115.9 mL/kg, respectively. These findings indicate that previous abdominal surgery may be a factor that causes more blood loss and blood product consumption as previously reported by other authors.22, 25

But not all Re-LT is associated with significantly more blood loss than primary LT. The time interval between primary LT and Re-LT in most GII patients was longer than 6 weeks (Table 1). There were two patients who underwent Re-LT within 3 weeks (5 days and 16 days, respectively) because of hepatic artery thrombosis. The blood loss during Re-LT was 13.5 mL/kg and 74 mL/kg, respectively. In contrast, two patients with hepatic artery thrombosis but underwent Re-LT beyond 6 weeks (44 days and 1072 days, respectively) had massive bleeding during Re-LT with blood loss of 336 mL/kg and 892 mL/kg, respectively. Primary LT per se, therefore, may not be the sole cause of more bleeding during Re-LT. The presence of adhesion in the great vessels and collateral vessels after primary LT is probably another contributing factor for increased bleeding during Re-LT. Because the number of our Re-LT series was small and only two patients underwent early Re-LT before adhesion in IVC or collateral vessels might have set in, our findings would require further validation with more cases over time.

5. Conclusion

Re-LT is associated with significantly more blood loss than primary LT. Significantly, more fluids and blood and blood product replacements are required. Greater amounts of sodium bicarbonate, calcium chloride, and neosynephrine are given in Re-LT patients to maintain hemodynamics. Our report suggests that more anesthetic personnel, more intravenous lines, and blood and blood products should be readily available when conducting anesthesia in Re-LT patients as these patients require more intensive fluid resuscitation and close monitoring because of increased perioperative bleeding.