New classification‐oriented treatment strategy for portal vein thrombosis after hepatectomy

Abstract Aim This study sought to evaluate the incidence, risk factors, and clinical outcomes of portal vein thrombosis after hepatectomy. Furthermore, we proposed a novel classification and treatment strategy for portal vein thrombosis after hepatectomy. Methods We retrospectively analyzed 398 patients who underwent hepatectomy and enhanced computed tomography imaging within 14 days after surgery in our hospital from 2009 to 2019. Portal vein thrombosis was classified into three categories according to the location of the thrombus – main, hilar, and peripheral – with main portal vein thrombosis further subclassified into three grades. Each patient's treatment strategy was determined based on their portal vein thrombosis classification and grading. From 2015, enhanced computed tomography imaging was performed routinely on patients who underwent anatomical hepatectomy on postoperative day 7. Results Portal vein thrombosis was diagnosed in 57 patients (14.3%) during the study period. Multivariate analysis revealed that a Pringle maneuver time of 75 minutes or longer was a significant predictor of portal vein thrombosis (P = .012). In total, 52 patients (91%) with portal vein thrombosis recovered by surgery, anticoagulant therapy, or without specific treatment. There was no instance of mortality recorded. Conclusions Patients who undergo hepatectomy are at high risk for portal vein thrombosis, especially when the Pringle maneuver time is long. The proposed classification and treatment strategy may be useful for clinical management of patients with portal vein thrombosis after hepatectomy.


| INTRODUC TI ON
Hepatectomy has been accepted as the standard treatment for patients with benign and primary liver cancer, as well as colorectal liver metastasis, with well-preserved liver function. 1 Despite recent developments in surgical techniques and perioperative management, postoperative complications may still lead to mortality. Portal vein thrombosis (PVT) is a frequent complication in cirrhosis; several factors are involved in PVT development and progression. 2 Similarly, PVT is a potentially life-threatening complication that also occurs after hepatobiliary surgery. PVT incidence after hepatic resection reportedly ranged from 2.1% to 9.1% in previous studies. [3][4][5][6] Postoperative PVT is typically asymptomatic but can cause acute intestinal ischemic necrosis, serious liver damage due to a lack of blood flow, and portal hypertension. This study, therefore, sought to evaluate the incidence, risk factors, treatment, and clinical outcomes of postoperative PVT after hepatectomy and proposed a novel classification and management scheme for addressing PVT after hepatectomy.

| Patients and methods
The study included 742 consecutive patients who underwent he- The clinical diagnosis of PVT after hepatectomy was confirmed using enhanced CT imaging. From January 2009 to December 2013, CT was conducted using a 16-detector-row CT machine, while, from January 2014 to December 2019, it was performed using a 128-detector-row CT machine. CT was performed with a 1.5-to 5-mm slice thickness. Enhanced CT was conducted on patients who presented with abdominal pain, fever, abnormal inflammation markers, liver enzyme levels, FDP levels, and D-dimer levels. Starting from 2015, enhanced CT was performed routinely in patients who underwent anatomical hepatectomy (≥ Couinaud 1 segment) on postoperative day (POD) 7.
All patients were tested routinely to discern the complete blood count, liver and kidney function, C-reactive protein level, and blood coagulation parameters, including PT activity, PT-INR, APTT, and fibrinogen level on preoperative day 1-3, PODs 1, 3, 5, and 7. Since 2015, serum FDP level, D-dimer level, and AT-III activity were routinely measured on preoperative day 1-3, PODs 1, 3, 5, and 7. We did not routinely use prophylactic anticoagulants postoperatively.
Postoperative variables included laboratory data, duration of hospital stay, reoperation for postoperative bleeding, 30-day mortality, and in-hospital mortality. The laboratory tests performed during the postoperative period were the same ones performed in the preoperative period. Postoperative bile leakage was defined as grade B/C according to the International Study Group of Liver Surgery. 8

| New classification of PVT
As part of this research, we propose a novel classification system for PVT after hepatectomy. PVT, in this study, was classified into the following three categories according to the site of thrombus ( Figure 1).

| Treatment strategy for PVT
For patients with PVT, the treatment strategy selected, according to the PVT grade, was as follows: We performed anticoagulation therapy using intravenous, unfractionated heparin for about 7 days, followed by warfarin or direct oral factor Xa inhibition for 3 months. AT-III concentrate was administered for more than 3 days when the serum AT-III activity was lower than 70%. Because anticoagulant treatment guidelines of PVT are not established, anticoagulation therapy did not involve specific anticoagulants, and various anticoagulants were irregularly given. We did not introduce anticoagulation therapy in patients who were at a high risk of bleeding. The target APTT was 1.5 times the normal value during heparin therapy, and the target PT-INR was 1.5 to 2.5 during warfarin therapy. Follow-up with enhanced CT took place after within 14 days; then the patients underwent enhanced CT imaging or enhanced magnetic resonance imaging within 6 months after the PVT diagnosis.
Resolved thrombus was defined as the complete or near complete resolution of the thrombus on imaging findings within 14 days or 6 months after the PVT diagnosis, while organized thrombus was defined as significant stenosis or occlusion of the vessels due to thrombosis on imaging findings at least 1 month after the PVT diagnosis.

| Surgical procedures
The extent of liver resection was based on the patient's functional liver reserve and the volumetric CT analysis results. The Pringle The portal vein flow was assessed by Doppler ultrasonography during reperfusion time. The hepatic parenchyma was divided, using a Cavitron ultrasonic surgical aspirator or the crush-clamping method, and each vessel encountered was ligated or clipped. For anatomical segmentectomy and subsegmentectomy, Glissonean pedicle transection or ultrasound-guided dye injection was performed as described previously. 9,10 The intrahepatic artery, portal vein, and hepatic vein flow were routinely assessed by Doppler ultrasonography after hepatic resection.

| Statistical analysis
Continuous variables are presented as means ± standard deviations or medians with ranges, while categorical variables are presented as frequencies and percentages. Univariate analysis was performed using Fisher's exact test or Welch's t test, as appropriate. Multivariate analysis was performed to identify significant predictive variables.
Receiver operating characteristic (ROC) curve analysis was performed for correlation analysis. The results were considered statistically significant when P-values were lower than .05. Statistical analyses were performed using SPSS version 20 software program (IBM Corp).

| RE SULTS
This study included 398 patients aged 18-90 years (mean age:  (Table 2). A comparison of the coagulation tests performed on preoperative day 1-3 and POD 1 between patients with and without PVT between 2009 and 2019 is shown in Table 3. The serum PT activity of patients with PVT on POD 1 was significantly lower than that of patients without PVT (P = .004).
According to the ROC curve analysis, a cutoff value of 61.5% for serum PT showed sensitivity of 42.6% and specificity of 80.7% (AUC = 0.618, P = .004).
Among the PVT patients, main thrombosis was found in 14 patients (25%), hilar thrombosis was found in 30 patients (53%), and peripheral thrombosis was found in 13 patients (23%). Of the patients with main thrombosis, grade 1 was noted in eight patients (14%), grade 2 was noted in four patients (7%), and grade 3 was noted in two patients (3%) ( Table 4). Of the patients with main thrombosis, three had both main and hilar thrombosis.
Out of the 57 patients with PVT, 24 patients (42%) were asymptomatic. PVT diagnosis was confirmed at a mean of six (range: [1][2][3][4][5][6][7][8][9][10][11] days postoperatively. In the PVT patients, the median duration of the postoperative hospital stay was 17 days, which was significantly longer than that of 13 days found among patients without PVT (P = .011).  Table 4. The complete clearance rate of PVT was 93% in the MPV, 80% in the hilar region, and 92% in the peripheral region, respectively. The treatment modalities and clinical outcomes are shown in Table 5.

| Treatments and outcomes
There was no instance of postoperative liver failure or in-hospital mortality among the patients with PVT in this study. Also, there was no case of postoperative bleeding that occurred among the patients treated with anticoagulant therapy.

| Correlation of the PVT location and operative procedures
Regarding the PVT location, main thrombosis showed a high rate of incidence in the right or left hemihepatectomy, while hilar thrombosis showed a high rate of incidence in anterior or posterior segmentectomy and left lateral or medial segmentectomy and subsegmentectomy (Table 6).

| D ISCUSS I ON
We found that the incidence of PVT after hepatectomy was 14.3%, and a duration of the Pringle maneuver of 75 minutes or longer is a significant predictive factor of PVT after hepatectomy. Furthermore, we proposed a novel classification system and treatment strategy for PVT after hepatectomy, which is easy to use clinically and provides a satisfactory outcome with a high rate of complete clearance of thrombosis.
The pathogenesis of thrombosis is based on Virchow's triadvenous stasis, hypercoagulable state, and endothelial injury 11 -and PVT is provoked by these three factors, which are interdependent and often coexist with one another. 12 A hypercoagulable state after hepatectomy has been reported as seen in this study, serum PT activity on POD 1 was lower in the PVT group than that in the control group. 13 The Pringle maneuver is a useful approach by which to control bleeding from the portal vein during hepatectomy, 14 but it can lead to stasis in the portal vein and endothelial injury. 12 In this study, a longer duration of the Pringle maneuver is a significant risk factor for PVT, which is consistent with previous reports. 3 Portal vein thrombosis can occur at any location, although it usually originates from surgically manipulated regions, such as the portal vein stump, and can be located only in the MPV due to kinking or compression. Therefore, we believe that developing a novel classification scheme for PVT after hepatectomy is necessary. In the present study, anticoagulant therapy was introduced immediately after PVT diagnosis, where appropriate, because of the possibility of thrombosis extension and decreased portal venous flow, which may result in liver failure or portal hypertension. In our treatment strategy for PVT after hepatectomy, the patients with main grade 2 or hilar thrombosis received anticoagulation therapy, and 75% and 80% of these individuals recovered from PVT, respectively. On the other hand, most of the patients with main grade 1 or patients with peripheral thrombosis did not receive anticoagulation therapy, and PVT resolved spontaneously in all these individuals, which is consistent with the findings of a previous study. 15 Of 30 patients with hilar thrombosis, six did not receive anticoagulation therapy because of a small thrombosis or refused treatment, yet the thrombosis resolved in all these individuals. This result suggests that small thrombi of the hilar region may be followed up for a short period without further action and, if seen to be worsening, anticoagulation therapy should be started immediately.
Incidentally, the location and etiology of thrombosis vary depending on the surgical procedure, which is supported by the result in Table 6. During hemi-hepatectomy, a thrombus is likely to be formed in the MPV; especially in right-sided hemi-hepatectomy, PVT is considered to be caused by the kinking of the PV trunk or turbulent flow in the PV, as reported previously. 3,5 During segmentectomy, a thrombus is likely to be formed in the hilar region, mostly in the stumps of the vessels. As reported by Mori et al, 15 gentle stretching and careful dissection around the PV may be necessary to prevent PVT. On the other hand, a thrombus located far away from the surgically manipulated region is considered to have formed in the MPV and, thereafter, flowed into hilar or peripheral region.
When PVT is detected in the early period after hepatectomy, the treatment effectiveness is likely to be higher than when it is detected later on after hepatectomy, 5 which is considered to be the tectomy is reported to be 56%. 5 Ultrasonography after surgery is noninvasive but shows difficulty in detecting PVT because of the abdominal incision and increased amount of bowel gas after surgery.
In contrast, the sensitivity of enhanced CT for PVT detection after hepatectomy was 100%. 5 In addition, Yoshiya et al 3 reported that PVT after hepatectomy is closely related with the delayed recovery of liver function and delayed liver regeneration. Therefore, making an accurate diagnosis using routine enhanced CT in the early period after hepatectomy is mandatory, and a rapid start of treatment is important.
The incidence of PVT after hepatectomy in this study was higher than that of previous studies. The higher incidence may be due to the selected patient study population because routine CT examination was performed in patients who underwent anatomical hepatectomy while excluding patients who underwent partial hepatectomy who may show a higher incidence of PVT. Because we focused on PVT during routine CT examinations, minimal or peripheral PVT were detected, and that may have been underestimated in other studies.
This study has several limitations. First, it was retrospective in nature and included patients who underwent enhanced CT postoperatively, while patients who did not undergo enhanced CT were excluded; thus, it has the potential of selection bias. Second, anticoagulation therapy did not involve specific anticoagulants, and various anticoagulants were irregularly given in this study, as anticoagulant treatment guidelines of PVT are not established. In the future, further studies are needed to establish the standard anticoagulant treatment agents for PVT. In addition, it is important to consider how to prevent postoperative PVT. Recently, an interesting study was reported that massage of hepatoduodenal ligament recovers portal vein flow immediately after the Pringle maneuver in hepatectomy. 16 In the future, we plan to study the improvement of the Pringle maneuver technique to prevent postoperative PVT.
In conclusion, patients who undergo hepatectomy are at high risk for PVT, especially when the duration of the Pringle maneuver is long. As many patients with PVT are asymptomatic and possibly progressive during a short period, routine CT screening is necessary in the early period after hepatectomy. The proposed classification and treatment strategy, based on the location and severity of the PVT, can help to improve clinical outcomes.