The friendly incidental portal vein thrombus in liver transplantation

Recent studies have shown that portal vein thrombosis (PVT) is feasible in liver transplantation (LT) compared to the past, where it was once considered a contraindication because of technical difficulty and was associated with a high mortality and morbidity. The early era of PVT management in LT showed significant mortality and morbidity. However, with improved surgical techniques, technical capabilities, and surgical innovations, we have seen a marked improvement in the management of PVT with lower mortality and morbidity as shown in recently published data.1, 2

The incidence of PVT has been reported to be approximately 2%-26% in various studies.3, 4 Therefore, a good preoperative plan and strategy for identifying patients undergoing LT with PVT are vital to ensure good perioperative outcome and long-term consequences. Despite careful preoperative planning with the various imaging techniques that are available, PVT is at times discovered incidentally during surgery and hence, may affect the operative strategy and outcome.5 Since there are limited data to date comparing the outcome of PVT when encountered incidentally during the intraoperative period with PVT diagnosed before surgery by preoperative imaging, this study aims to compare the overall outcomes of patients with PVT in these 2 groups based on data from a center with decades of experience in LT.

PATIENTS AND METHODS

RESULTS

Survival Analysis of Graft and Patient Survival

The median graft survival times were 41.3 (0-75.8), 38.1 (5.0-75.0), and 33.6 (12.3-75.5) months in group 0, group 1, and group 2, respectively (P = 0.65). The 1-, 3-, and 5-year overall graft survival rates were 91.3%, 88.2%, and 86.8% in group 0; 96.4%, 84.6%, and 84.6% in group 1; and 96.9%, 91.8%, and 91.8% in group 2, respectively (Fig. 1).

The median patient survival times were 41.3 (range, 0-75.8), 38.1 (range, 5.0-75.0), and 33.6 (range, 12.3-75.5) months in group 0, group 1, and group 2, respectively (P = 0.66). The 1-, 3-, and 5-year overall recipient survival rates were 92.0%, 88.8%, and 87.4% in group 0; 96.4%, 84.6%, and 84.6% in group 1; and 96.9%, 91.8%, and 91.8% in group 2, respectively (Fig. 2).

DISCUSSION

PVT in LT has always been a challenge, but it is no longer considered a hindrance to LT since the first reported case of PVT in LT by Shaw et al.6 in 1985. Our experience with patients with PVT having LTs in this center showed that the incidence of PVTs was 15.7%, which was similar to the reported incidence worldwide.2

The study also showed that all the groups had similar preoperative characteristics except for their MELD scores, tumor diagnoses, preoperative platelet levels, serum bilirubin levels, and waiting times, especially for those undergoing DDLT.

With regard to tumor diagnosis, a higher incidence was seen in the incidental PVT group. The presence of PVT in HCC is generally from clots due to a procoagulable state from HCC, which can be easily removed by thrombectomy as seen mainly in the incidental group. Evidence of HCC with vascular involvement would have been contraindicated for LT in our center. The tumors that were found were generally peripheral and not close to any vessel.

In this series, it was also found that the waiting time to LT especially in DDLT showed a shorter median time for the non-PVT group. We have found that the non-PVT group contained a higher number of patients with acute fulminant liver failure, whereas there was none in both PVT groups; hence, these patients did not wait long for LT because a graft was provided as a result of a lifesaving situation where some patients were not part of the LT waiting list. With PVT, both group 1 and group 2 had comparable waiting times, but the population in this group was smaller in numbers than the non-PVT group.

Only a high serum bilirubin level was found to be significant on multivariate analysis when an analysis for the risk factors for developing PVT was made. However, this finding is not consistent or is paradoxical in our series because the non-PVT group was noted to show a higher median score than both PVT groups. Most likely, the role of serum bilirubin is not clinically useful in predicting PVT. The finding of more patients with acute fulminant liver failure in the non-PVT group and none in the PVT group probably contributed to a higher preoperative serum bilirubin level in the non-PVT group. Our risk factor differs with some studies, which showed age, sex, MELD score, Child-Pugh score, liver cirrhosis, or low platelet counts as a risk for PVT.2, 7 Another potential risk factor for PVT development was the association with interventions used for the treatment of portal hypertension such as splenectomy, shunt operation, transjugular intrahepatic portosystemic shunts, and sclerotherapy for varices but none of our patients had undergone similar interventions; hence, PVT could not be evaluated in this aspect.3

This study was able to show that the outcome in OS in patients and grafts was similar to patients without PVT. Before the year 2000, most literature showed that despite LT in PVT being a possibility, the perioperative mortality was high and the 1-year survival was below 80%, which differed from our recent findings.8-10 With recent advances and knowledge of the disease as well as with improved surgical techniques, the outcomes and survival in patents with PVT undergoing LT have seen improvement as is evident in this series as well.11 However, Fouzas et al.12 noted that outcome and survival remained poor despite the technical innovations but that it was related to the grade of PVT instead.

In this study, further investigations and comparisons on the basis that PVTs were identified preoperatively (group 1) or incidental (group 2) were made, of which there was a scarcity of data comparing both situations. It is imperative that careful preoperative imaging to identify PVT before LT is determined when deciding the surgical strategy to be employed during the operation.13, 14 PVT discovered intraoperatively might increase the complexity of the surgery. The incidence of incidental PVT was reported to be 12%-64%, and this study revealed that incidental PVT accounted for 51.7% in our center.5, 7, 15

Early imaging before transplantation might be able to identify PVT, but as shown in this study, it bears no significance as incidental PVTs were still encountered intraoperatively despite imaging being done less than 1 month before surgery.

However, preoperative imaging still plays a role in determining the surgical strategy to be employed. If PVT was encountered before LT especially for those with Yerdel grades 3 or 4, this group of patients was in need of more advanced surgical techniques rather than a conventional thrombectomy, as the study had shown in instances where thromboendovenectomy, coronary vein anastomosis, and SMV graft to SMV native anastomosis were performed for patients at this grade. We could also safely determine that if preoperative imaging did not show PVT but PVT was discovered incidentally during surgery, then the PVT might not be as extensive as PVT is more likely of Yerdel grades 1 or 2; hence, a conventional thrombectomy or occasionally thromboendovenectomy would suffice, which would not require a highly complex intervention.

Although this study did not describe in length which of the various surgical strategies to use in relation to the various Yerdel grades encountered, for grade 1 or 2, a conventional thrombectomy or thromboendovenectomy is usually adequate.11, 16, 17 Occasionally, a resection of an affected short segment with primary anastomosis of the portal vein can be performed as well.7 For Yerdel grades 3 or 4, a more extensive and technical surgical strategy is often needed. Although we have described our techniques used for Yerdel grades 3 or 4, other sources in the literature have also reported various techniques, such as an interposition graft using a native or jump graft, renoportal anastomosis, portocaval hemitransposition, portal arterialization, or multivisceral transplant.15, 18-22

In the case of LDLT, the donor graft usually has a short portal vein end, which may present further challenge in the event of a recipient having a PVT especially when a simple thrombectomy or thromboendovenectomy cannot be performed. In such instance, a few approaches have been described to overcome this, such as recipient portal vein plasty, portal vein stenting, interposition graft use, or anastomosis to venous collaterals, where inflow anastomosis using paracholedochal veins had been described.23-25 Although they did not classify PVT according to Yerdel classification in their series of PVT in LDLT, Egawa et al.26 described techniques to overcome extensive PVT involving the portal trunk with SMV or splenic vein involvement where the “pull-through technique” was applied with subsequent eversion or incision technique. However, when the “pull-through technique” was not feasible, the SMV can be transected with an interposition graft using the recipient's left portal vein in right lobe LDLT.

Unlike LDLT, in DDLT, the long donor PV in DDLT gives surgeons an added benefit for the ease of portal vein anastomosis during the recipient's surgery. However, this study reports the discovery of PVT during the recipient LT where most PVs are isolated and transected close to the liver hilum with thrombosis being discovered and removed by a conventional thrombectomy. Having a long PV from the donor would probably allow a short segment of the recipient's PV to be transected at any point where it was deemed unhealthy.

When the postoperative outcomes in all 3 groups were compared, they were comparatively similar as well. Of interest would be the early postoperative portal vein rethrombosis where both groups 1 and 2 had slightly higher incidences than the non-PVT group although it remained <5%. This is less than most published studies, which have a reported incidence of 6%-40%.9, 27-29 In our series, we do not practice the use of anticoagulants in either the preoperative or postoperative period in our patients with PVT going for LT. Although several published series on the use of anticoagulants in PVT have been published to date, there is still no sufficient evidence for its use in clinical practice or as part of the standard guidelines.20, 30 The need to weigh between the risk of bleeding and a procoagulable state in patients with PVT for LT should be considered carefully before deciding to start such patients on anticoagulants as more evidence or randomized controlled trials are still required.

Our late complication also showed a high incidence of portal vein stenosis in group 1, which can be attributed to the group having a number of subjects with a higher Yerdel grade. However, in this series, despite reports of portal vein stenosis based on radiological imaging, very few patients required interventions, such as an angioplasty or reconstruction, where only 1 patient from either group had angioplasty with the rest requiring close follow-up as good venous patency and graft function continued throughout their follow-up.

This study has a limitation where we encountered difficulty in stratifying PVT with Yerdel grades 1 or 2 in the group where PVT was discovered incidentally during surgery. This was because it was difficult to differentiate between either Yerdel grades intraoperatively.

We hope that the study conducted would provide clinicians with further information for the management of PVT especially when incidental PVT is involved. An algorithm with regards to the management of PVT inclusive of incidental PVT could be determined as shown in Fig. 3 on the basis of the strategies employed in our institution, with further additional strategies for Yerdel grades 3 or 4 being described in other literature.26

In conclusion, the study has shown that patients with PVT undergoing LT can have similar postoperative outcomes and OS as compared to those with non-PVT including instances of where PVT is discovered incidentally during the intraoperative period. The discovery of incidental PVT only requires thrombectomy, and it added little to the complexity of the transplantation where the outcome was not adversely affected because most will be of a lower Yerdel grade. Preoperative imaging is very useful to identify those with a higher Yerdel grade in order to allow proper planning of the surgical strategy during transplantation. With recent advances, the outcome of patients with PVT has improved in the last 5 years as shown by this data.