Favorable Outcomes After Liver Transplantation With Normothermic Regional Perfusion From Donors After Circulatory Death: A Single-center Experience

Abstract

Background. 

Controlled donation after circulatory death (cDCD) has been associated with a high incidence of ischemic cholangiopathy and other perioperative complications. In an attempt to avoid these complications, we implemented an active protocol of cDCD liver transplant (LT) with normothermic regional perfusion (NRP) preservation.

Methods. 

This is a descriptive analysis of data collected from a prospective date base of cDCD LT preserved with NRP from January 2015 to June 2017 with a minimum follow up of 9 months.

Results. 

Fifty-seven potential cDCD donors were connected to the NRP system. Of these, 46 livers were transplanted over a 30-month period (80% liver recovery rate). The median posttransplant peak in alanine transaminase was 1136 U/L (220-6683 U/L). Seven (15%) patients presented postreperfusion syndrome and 11 (23%) showed early allograft dysfunction. No cases of ischemic cholangiopathy were diagnosed, and no graft loss was observed over a medium follow-up period of 19 months. Of note, 13 donors were older than 65 years, achieving comparable perioperative and midterm results to younger donors.

Conclusions. 

As far as we know, this represents the largest published series of cDCD LT with NRP preservation. Our results demonstrate that cDCD liver grafts preserved with NRP appear far superior to those obtained by the conventional rapid recovery technique.

Despite increasing transplant activity, there is still a significant imbalance between supply and demand in liver transplantation (LT), which has led to the use of so-called marginal donors to expand the donor pool.

Controlled donation after circulatory death (cDCD) represents an important source of liver grafts in countries with an appropriate legal framework, accounting for 27% of LT donors in the Netherlands and 18% in the United Kingdom.¹ However, cDCD has been associated with a high incidence of ischemic cholangiopathy (IC), reported in up to 34% of cases,^2-4 as well as other perioperative complications, such as primary nonfunction (PNF), hepatic arterial thrombosis (HAT), and acute kidney injury (AKI).^5-8 These apparently poor results have been attributed to the unavoidable warm ischemia time (WIT), which enhances organ preservation injury. As a result, according to the donor risk index, cDCD status is considered an independent factor, strongly associated with graft loss and therefore with the need of retransplantation or patient death.^3,5,9 Normothermic regional perfusion (NRP) has been reported to decrease the incidence of IC in cDCD LT.^10-12 Unlike other countries, Spanish donation law permits the premortem cannulation (PMC) of potential donors;¹³ thereby significantly reducing WIT as urgent cannulation is not required. Based on these facts, we implemented an active cDCD LT program with PMC and NRP.

Herein, we present and analyze the experience and preliminary outcomes of this program, which is, to our knowledge, the largest series of cDCD LT with PMC and NRP. We also discuss the advantages of NRP when selecting good quality cDCD liver grafts.

MATERIALS AND METHODS

Study Design

We performed a descriptive analysis of donor, recipient, and operative variables collected from a prospective database of cDCD LT with NRP. The study was conducted from January 2015 to June 2017, and all patients had a minimum follow-up period of 9 months. All transplants were performed at the same hospital, although liver grafts were retrieved at 5 different regional centers. All grafts were obtained using standardized NRP equipment according to the institutional protocol, performed by the same surgical team which moved to the donor hospitals. The PMC procedure, withdrawal of life-sustaining treatment, donation, acceptance by next of kin, and signing of informed consent were conducted according to Spanish donation law.¹³ This study was approved by the institutional review board.

The donor selection criteria were the same as for donation after brain death (DBD). Although an initial age limit of 65 years was established, from the tenth patient onwards we did not consider a specific limit but tried to avoid multiple associated comorbidities (AC) such as hypertension, diabetes, dyslipidemia or peripheral vasculopathy in older donors. Considering this protocol modification, we performed a comparative analysis of outcomes between donors aged under and over 65 years.

Procurement, Preservation, and Transplantation Protocols

Our NRP protocol was similar to that published by Miñambres et al.¹² In brief, PMC was performed in an interventional radiology room 2-3 hours before withdrawal of life-sustaining treatment. The donor was then returned to the intensive care unit, where catheters were primed and connected to the NRP system. After extubation, the donor was heparinized (3 mg/kg), and blood pressure was observed. Upon confirmation of death by an independent physician and after a 5-minute standoff period, the aortic balloon was inflated, absence of radial flow in the arterial catheter was confirmed, and the NRP system was initiated. The WIT was considered the time from when the systolic blood pressure reached lower than 60 mm Hg until the time at which NRP perfusion was initiated, including 5-minute nontouch period. We only accepted livers with a WIT of <30 minutes. The pump flow kept above 1.7 L/min, perfusion pressure above 60 mm Hg, pH between 7.35 and 7.45, hematocrit above 20%, and temperature stable at 37°C. Serum transaminase and lactate levels were monitored every 20 minutes. According to protocol, the liver was discarded if the transaminase level exceeded >3 times the normal level at initiation or 4 times this level at the end of NRP. Afterward, we transferred the donor, with continuous NRP continuing, to the operating room where a laparotomy was performed to assess the quality of the liver and bile duct, which was sectioned in order to evaluate its vascularization and macroscopic aspect. Finally, after approximately 2 hours of NRP, a standard donor procurement surgery, including a dual cold organ perfusion, was performed (4°C; Celsior) and cold-stored.

Organ allocations were based on the model for end stage liver disease (MELD) system and Child-Pugh score and were the same as for DBD LT. As with our conventional recipients, we performed a piggy-back technique without portacaval shunt and an end-to-end choledochocholedochostomy with T tube for biliary reconstruction. A routine cholangiogram was performed at 1 week and 3 months posttransplantation before elective T tube removal,¹⁴ and a Doppler ultrasound exploration was conducted within the first 48 hours, at 1 week and 1 and 3 months after LT. The immunosuppressive therapy involved basiliximab (Simulect), mycophenolate mofetil, and steroids with a delayed introduction of extended-release tacrolimus (Advagraf).

Endpoints

The primary endpoint was graft survival. The secondary included the development of biliary complications (anastomotic strictures or leaks and IC), patient survival, incidence of postreperfusion syndrome (PRS),¹⁵ posttransplant alanine transaminase (ALT) peak (highest level within the first 7 days), and early allograft dysfunction (EAD) rate.

Postreperfusion syndrome was considered when the mean arterial blood pressure was 30% lower than the previous value within 5 minutes of unclamping, which lasted for at least 1 minute. Early allograft dysfunction was defined according Olthoff’s criteria,¹⁶ and IC was defined as the presence of nonanastomotic biliary strictures without a concomitant hepatic artery thrombosis.^4,17,18

Statistical Analysis

Qualitative variables are presented using the number and percentage and quantitative variables using the median and range. Comparisons between groups were performed using Fisher and Mann-Whitney U tests for qualitative and quantitative variables, respectively. The significance threshold was P <0.05. The patient and graft survival data were analyzed by the Kaplan-Meier method. Analyses were conducted using Stata software (version 14.2; StataCorp, College Station, TX).

RESULTS

A total of 215 deceased LT were performed at our center during the studied period, 169 of which were DBD LT and 46 DCD LT, comprising the study group. Fifty-seven potential cDCD donors were identified and connected to the NRP system. Of these, 11 grafts were discarded, 6 due to suboptimal macroscopic aspect, 2 exceeded the ALT limit during NRP, 1 due to a kidney tumor, 1 due to bilateral hepatolithiasis, and 1 caused by a technical failure of the pump. Therefore, 46 livers were transplanted, representing a liver recovery rate of 80%. Donor and recipient demographics are presented in Table 1. Although the median donor age was 58 (27-76) years, 13 donors were older than 65 years (7 of them older than 70 y). Seven liver grafts (15%) had mild macrovesicular hepatic steatosis (<30%). cDCD grafts were also implanted in 2 combined liver-kidney transplantations, 1 elective retransplantation, and 1 urgent 2-stage liver retransplantation. The median MELD score was 12 (7-27), but in the subgroup of 4 (8.6%) patients who developed AKI¹⁹ (peak serum creatinine ≥2 times the baseline level) was 25 (20-27). The transaminase level remained stable during NRP preservation, with median ALT and AST levels of 27.5 U/dL (5-84 U/dL) and 32 U/dL (6-152 U/dL), respectively, at the beginning, and 30.5 U/dL (7-116 U/dL) and 34.5 U/dL (9-171 U/dL), respectively, at the end of the NRP preservation. The lactate levels decreased from a median level of 40 mg/dL (5-140 U/dL) at the beginning of NRP to 17 mg/dL (7-58 U/dL) at the end (Figure 1). As shown in Table 2, the median duration of WIT, NRP, and cold ischemia time (CIT) were 10 minutes (6-22 min), 126.5 minutes (86-161 min), and 284.5 minutes (149-409 min), respectively. Seven (15%) patients developed PRS and 6 (13%) showed fibrinolytic syndrome detected by thromboelastography. The median ALT peak was 1136 U/L (220-6683 U/L), and 11 (23%) patients developed EAD. Neither retransplantation recipients nor the combined liver-kidney transplantations needed renal replacement therapy, which was only required in a patient with a severe PRS. Extended release tacrolimus was introduced at a median of 4 days (1-22 days) after LT. The median intensive care unit and hospital stay were 4 days (1-24 days) and 14 days (7–55 days), respectively. Although there were more donor ACs for the group with donors aged over 65 years (Table 1), outcomes between donors aged over and under 65 were comparable (Table 2). Over a median follow-up of 19 (9-40) months, no grafts were lost and all patients remained alive, with neither clinical nor radiological evidence of IC. One patient developed an asymptomatic biliary anastomotic stricture and 4 (8.7%) cases of arterial stenosis were successfully treated with stents during the postoperative period.

TABLE 1.: Donor and recipient demographics, categorized by donor age

TABLE 2.: Preservation times and overall perioperative and midterm outcomes, categorized by donor age

FIGURE 1.: Evolution of analytical parameters during NRP (median). ALT, alanine transaminase; NRP, normothermic regional perfusion.

DISCUSSION

The use of liver grafts from cDCD donors remains controversial because of the risk of graft failure and biliary complications. In an attempt to avoid these complications, we implemented a PMC and NRP protocol to reduce WIT and optimize graft selection. In our series of 46 consecutive liver transplants over 30 months, we achieved comparable perioperative and midterm results to our series of DBD LT²⁰ without any graft loss and with an outstanding 80% liver recovery rate, similar to that obtained with DBD donors in Spain.²¹ Noteworthy, over a medium follow-up period of 19 months, only 1 patient presented an asymptomatic biliary stenosis, and no cases of IC were diagnosed. We transplanted 13 livers from cDCD donors older than 65 years. At the beginning of the study, donor age was limited to 65 years; however, given the initially good perioperative results, in terms of graft function, we decided to cross this age threshold, achieving the same perioperative and midterm results as younger donors. Although some authors have recently reported good experiences with donors older than 60 years,²² donor age still remains as a limiting factor in cDCD LT results. In fact, in a recent multicenter analysis, authors reported greater risk for biliary complications and IC with cDCD donors older than 50 years.²³ This contrasts with our excellent results, and it could be justified by the use of NRP. Ischemic cholangiopathy is a characteristic complication of cDCD LT and has been linked to arteriolonecrosis and microthrombi formation during warm ischemia and graft preservation.^7,24 Indeed, some authors have reported improved results by shortening the ischemia time and additional flushing of the graft with thrombolytic agents such as tissue plasminogen activator or urokinase.^7,25,26 However, the use of NRP for liver preservation, by itself, seems to limit arteriolar and biliary necrosis,²⁷ thus reducing postoperatory biliary complications. Predictably, donors aged over 65 years had a significantly higher number of ACs; however, it seemed to have little impact on the outcomes, probably due to the ability of NRP to differentiate poor from adequate grafts. Our favorable results might be partly related to the recipients’ low MELD score. In our center, we use the laboratory MELD score without additional MELD exception points, but recipients are also evaluated by Child-Pugh score, so refractory ascites, encephalopathy, and other clinical situations can be prioritized. In addition, due to a high donation rate, the median waiting time for LT is <3 months, thus avoiding the deterioration of the patient.

Higher incidence of HAT has been already described after cDCD LT.⁶ We also observed a relatively high rate of arterial stenosis (8.7%) that did not evolve to HAT. In our experience, protocol Doppler follow-up permitted early detection of these complications, which were treated with stents, thereby avoiding progression to HAT. Of note, all cases of stenosis occurred during the first 10 transplants and, to date, we have performed >55 cDCD LT with no further arterial complications, even though we have not changed any technical aspects (data not shown).

The incidence of EAD, PNF, and AKI in our recipients was not different to what has been reported for DBD LT.²⁰ These results are in contrast with other reports in which cDCD grafts were associated with a high risk of perioperative complications, such as EAD, PNF, and AKI, which significantly impacted graft survival.^2-8 In our series, only 4 patients developed postoperative AKI (8.6%) coinciding with EAD and high ALT peak. Leithead et al⁸ reported an increased frequency of AKI with cDCD donors and a correlation between AKI and the AST peak as a marker of ischemia-reperfusion injury. Therefore, the low incidence of AKI and low ALT peak makes us believe that the use of NRP is associated with the decrease of ischemia-reperfusion injury. Some experimental studies have evidenced that the recirculation of warm, oxygenated blood significantly improves the viability of livers previously subjected to warm ischemia, decreasing ischemia-reperfusion injury.^27,28 Normothermic regional perfusion has also shown improvements in clinical results after uncontrolled DCD LT,²⁹ and more recently, application of NRP in cDCD has also been associated with improved outcomes. Rojas-Peña et al¹⁰ performed 13 cDCD LT with 1 case each of PNF and 1 of IC and a 1-year graft survival rate of 85%. Similarly, Oniscu et al¹¹ and Miñambres et al¹² reported results from different cohorts of 11 cDCD LT with no evidence of IC, although 1 patient in each series suffered PNF and died as a result. With a median follow-up period of 19 months, our results support these experiences with no cases of IC or PNF and 100% graft/patient survival.

In our opinion, NRP offers multiple benefits for the recovery and evaluation of suitable grafts. First, NRP restores blood flow to the liver before cold storage in a way that is similar to the ischemic preconditioning techniques used during liver resection.²⁸ Second, NRP permits real-time monitoring of transaminase and lactate levels while maintaining the donor in a more physiological situation. This allows us to anticipate and discriminate the livers, depending on the degree of ischemic injury and recovery capacity of each one. In our experience, lactate levels, which increase due to anaerobic metabolism during cardiac arrest,³⁰ were found to be decreased during NRP, reflecting adequate organ perfusion and clearance capacity of the functioning liver. An unexpected progressive elevation in lactate without justifiable cause could be considered an indicator of poor organ perfusion and therefore a reason to refuse the liver. Finally, NRP permits a more relaxed laparotomy favoring proper assessment of hepatic perfusion and the bile duct.

Premortem cannulation is another key factor for our good results. Premortem cannulation reduces WIT, helping to attenuate the ischemic injury. Premortem cannulation was always performed by a specialized radiologist in an interventional radiology room, allowing us to identify anatomic anomalies, confirm the correct placement of the balloon, and estimate the filling volume of the balloon required to block the aorta. These data were always registered to guarantee adequate closure of the thoracic aorta and it is confirmed by the absence of radial artery pulse, thereby avoiding unwanted donor resuscitation. We understand that PMC can raise obvious ethical and legal concerns that make it nonapplicable to some countries. Spanish law permits the cannulation and heparinization of donors before treatment withdrawal,¹³ whereas in many other countries, premortem interventions are prohibited. However, it is possible to benefit from the NRP without PMC. Oniscu et al¹² described a novel procedure with postmortem cannulation and subsequent connection to NRP, showing good posttransplant outcomes and higher liver utilization rates when compared with the super rapid recovery procedure.

Our study has several limitations, such as the small number of patients and absence of a control group; however, it included a homogeneous cohort and was carried out using an easily reproducible technique. On the other hand, this is the largest published series of cDCD LT preserved with NRP, with the strengths being a significant follow-up period and similar results to DBD in terms of survival and complications. Based on our excellent results, we do not consider the need for a control group with the super rapid recovery technique.

In summary, we believe that NRP is a valuable technique to achieve quality cDCD grafts irrespective of donor age, therefore allowing expansion of the donor pool. We do not consider cDCD livers to be marginal grafts, per se, and its use is a well-established procedure in our transplant system contributing with >30% of all transplanted hepatic grafts. Although more extensive multicenter studies are required, our excellent results and the recently published results of others demonstrate that cDCD liver grafts preserved with NRP are superior to those obtained with the conventional rapid recovery technique and even comparable to DBD grafts.

ACKNOWLEDGMENTS

The authors thank to Hospital Universitario Cruces (Barakaldo), Hospital Universitario Donostia (Donostia), Hospital Universitario de Araba (Vitoria), and Hospital Universitario de Basurto (Bilbao) for their collaboration in the DCD donation process.