I read with great interest the recent article, “Liver transplant in patients with portal vein thrombosis: the experience of 55 patients,” published by Ozer and colleagues.1 The authors stated that liver transplant can be performed successfully with appropriate surgical techniques, regardless of the stage of portal vein thrombosis (PVT). Herein, I share my opinion about this valuable work.
Liver cirrhosis is a known risk factor for PVT development regardless of the underlying cause, and this risk is directly related to the duration and severity of liver cirrhosis. Also, it has been established that the presence of hepatocellular carcinoma (HCC) in a cirrhotic liver increases the risk of PVT. The authors of this article reported that 23.6% of the patients presented in this study had HCC, which was higher than the HCC incidence rates found in high-volume centers.2,3 In my opinion, the authors could have easily determined whether HCC was an independent risk factor for PVT by comparing HCC incidence rates in patients with PVT versus patients without PVT, which would be an important finding for readers.
Ozer and colleagues stated in the first paragraph of the results section that “cirrhosis, secondary to hepatitis B virus infection (16%), was the most frequently detected primary reason for transplant; 22 patients (40%) had cryptogenic cirrhosis,” but this ratio is not correlated with the indications for liver transplant in Turkey. A careful examination of Table 1 showed that 29.1% of patients with PVT underwent liver transplant to treat liver disease due to hepatitis B virus. The 16% rate used in this section does not match the results of any data presented by the authors.
Ozer and colleagues stated that they used portal vein arterialization to provide sufficient portal flow in a patient with an extra-anatomic anastomosis. As previously established, there are serious complications related to portal vein arterialization, including portal hypertension secondary to high arterial flow, right heart failure, hepatocyte damage, and fibrosis in the graft. Therefore, in my opinion, portal vein arterialization should be avoided whenever possible. The fact that the patient presented by the authors died in a very short time despite retransplant supports this opinion. If sufficient portal inflow is not provided, then cavoportal hemitransposition or renoportal anastomosis may be used as an alternative technique. According to the large case series published by Bhangui and colleagues, satisfactory results were obtained especially with renoportal anastomosis.4
Ozer and colleagues reported that patients with PVT received low-molecular-weight heparin in the early posttransplant period and were then discharged with acetylsalicylic acid treatment and that 2 patients were given anticoagulant therapy because Budd-Chiari syndrome was detected. I disagree with the algorithms of coagulation prophylaxis used by the authors. Although liver cirrhosis is an important cause of PVT, a considerable proportion of PVT patients have an underlying hematological problem or idiopathic disorder. Therefore, even when the surgical technique is normal in these patients, it remains difficult to predict the course of the thrombotic process in the early postoperative period. Therefore, at least 1 or more anticoagulant and antiplatelet therapy option should be given in patients with PVT for at least 3 to 6 months, especially in patients in a hypercoagulable state or with nonphysiological portal inflow (eg, vascular graft, extra-anatomic anastomosis).5-7
Ozer and colleagues used the Kaplan-Meier test and the chi-square test interchangeably, which caused a serious mistake. The authors applied the chi-square test to calculate whether there was a significant difference in mortality between the groups with and without PVT. However, the P value published by the authors is incorrect. With the licensed edition of the SPSS software (version 25), comparison of the data for the 2 groups shows P = .072 with the Yates-corrected analysis The value of P = .05 published by the authors should be corrected to properly represent the comparison; that is, there was no significant difference in mortality between the groups with and without PVT. To compare the groups with the chi-square test, it does not matter how long the patients survived. On the other hand, the Kaplan-Meier curves and estimates test is the best test to calculate the probability of patient survival. Therefore, new statistical analyses are required to produce a new graph of the Kaplan-Meier test, along with a new corrected P value, for Figure 1.
A similar mistake was made in the second sentence of the last paragraph of the results section: the authors calculated P = .29 for the comparison of mortality between extra-anatomic and anatomic portal vein anastomosis. However, according to the Fisher exact test, the statistical value should be P = .236. The error (P = .29) should be corrected. Likewise, a new Kaplan-Meier curves and estimates graph should be presented for Figure 2, along with a corrected P value, for the comparison of the cumulative survival between these 2 groups.
DOI : 10.6002/ect.2019.0382
From the Department of Surgery and Liver Transplant Institute, Inonu University Faculty of Medicine, Malatya, Turkey
Acknowledgements: The author has not received any funding or grants in support of the presented research or for the preparation of this work and has no declarations of potential conflicts of interest.
Corresponding author: Sami Akbulut, Department of Surgery and Liver Transplant Institute, Inonu University Faculty of Medicine, Elazig Yolu 10, Km Malatya 44280, Turkey
Phone: +90 422 341 0660