Acute splanchnic vein thrombosis in patients with COVID-19: A systematic review

There is increasing evidence that coronavirus disease 2019 (COVID-19) is associated with a significant risk of venous thromboembolism. While information are mainly available for deep vein thrombosis of the lower limb and pulmonary embolism, scarce data exist regarding acute splanchnic vein thrombosis (SVT) in this setting. PubMed, EMBASE and Google Scholar English-language articles published up to 30 January 2021 on SVT in COVID-19 were searched. Overall, 21 articles reporting equal number of patients were identified. 15 subjects presented with portal vein thrombosis, 11 with mesenteric vein thrombosis, four with splenic vein thrombosis, and two with Budd-Chiari syndrome. Male sex was prevalent (15 patients), and median age was 43 years (range 26–79 years). Three patients had a history of liver disease, while no subject had known myeloproliferative syndrome. Clinical presentation included mainly gastrointestinal symptoms. Anticoagulation was started in 16 patients. Three patients underwent bowel resection. Ten subjects developed gastric or bowel ischemia, seven of whom underwent bowel resection, and four died after SVT diagnosis. Although rare, SVT should be seen as a complication of COVID-19. Patients with severe gastrointestinal symptoms should be screened for SVT, as rapid recognition and correct management are essential to improve the outcome of these patients.


Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is identified as responsible of coronavirus disease 2019 (COVID- 19), a clinical condition ranging from mild symptoms, such as impairment of the smell and taste, to the typical pulmonary manifestations including acute respiratory distress syndrome [1] . The virus may also directly damage the intestinal mucosa [2] , as gastrointestinal symptoms have been widely reported [3 , 4] . Furthermore, increasing evidence showed that COVID-19 might be associated with hemostasis impairment, thus predisposing patients to both venous and arterial thromboembolism [5] . While the association between COVID-19 and deep vein thrombosis of the lower limb and pulmonary embolism has been extensively investigated, less is known about thrombotic events in other districts, such as splanchnic vein thrombosis (SVT). SVT is an uncommon manifes-tation of venous thromboembolism (VTE) that includes portal vein thrombosis (PVT), mesenteric vein thrombosis (MVT), and thrombosis of the liver veins or vena cava inferior (Budd-Chiari syndrome, BCS). SVT is more often diagnosed in patients with cirrhosis or liver malignancy. Other predisposing factors, which are often observed in patients with non-cirrhotic liver, are intra-abdominal surgery, infections and inflammatory diseases, as well as inherited or acquired thrombophilia [6] .
Assuming that the splanchnic venous system could be also affected by the COVID-19-related coagulopathy, we conducted a systematic review of current literature on SVT and COVID-19.

Methods
We searched in PubMed, EMBASE and Google Scholar Englishlanguage articles published between November 2019 and 30th January 2021 including the following MeSH terms "COVID- 19  sis", "COVID-19 and Budd-Chiari syndrome", "COVID-19 and bowel ischemia". Case reports, case series, commentaries, letters to editors, and review articles were considered. We evaluated independently title and abstract for inclusion, double-checking for duplication and extracting the data using the text, tables, and figures of the original published articles. Cross-referencing yielded no additional records and duplicate articles were assessed and subsequently eliminated. All the demographic and clinical characteristics, clinical course and outcomes are reported using median and incidence.

Results
Overall, 641 articles were screened, 21 of which were selected according to the above mentioned criteria [7 -27] . The included articles were all case reports (except for one case series [10] ), and contained description of 21 cases. Overall, 15 patients presented with PVT, 11 with MVT, 4 with splenic vein thrombosis, and 2 with BCS. The clinical features, management and outcome are showed in Tables 1 and 2 and summarized in Fig. 2 .
Comorbidities included cardiovascular risk factors like obesity (3 patients), type 2 diabetes (2 patients), arterial hypertension and hyperlipidemia (one patient each), as well as asthma (2 patients), Parkinson disease and vascular dementia (one patient each). Three patients had a history of known liver disease (non-alcoholic steatohepatitis (NASH) plus hepatitis B virus (HBV) cirrhosis, chronic hepatitis B, and alcohol-associated cirrhosis). Except for one patient with operated craniopharyngioma, none of the patients had a history of neoplasia or myeloproliferative neoplasms (MPN). In one case, imaging tests performed at admission documented multifocal liver lesions consistent with hepatocellular carcinoma in the setting of newly discovered chronic hepatitis B, whereas no neoplastic thrombus was reported in the patient with known cirrhosis on NASH and HBV on the investigations performed. For ten patients, no comorbidities were reported.

Presentation
Symptoms and/or signs of SVT were present at COVID-19 onset in ten cases, whereas in seven patients SVT represented a belated manifestation (range 5-21 days). For one subject, SVT was diagnosed as occasional finding with imaging performed in the setting of COVID-19.
Clinical presentation included mainly gastrointestinal symptoms, such as abdominal pain (16 patients), and vomiting (7 patients). Jaundice and diarrhea were reported in one case each. Fever was present for 7 patients, while dyspnea, cough and altered mental status were reported in 4 cases each. One patient had a large volume hematemesis following initiation of heparin for a lower extremity deep vein thrombosis during COVID-19.

Diagnosis of SVT and of COVID-19
SVT was diagnosed in the outpatient setting in 17 cases. Almost all the diagnosis of SVT required a computed tomography (CT) scan to be confirmed, except in one case (abdominal Doppler ultrasound, DUS). Eight patients displayed multivessel involvement, including upper mesenteric artery for one patient. At the time of SVT diagnosis, features of gastric or bowel ischemia were reported in ten cases.
Diagnostic tests for COVID-19 included real-time reverse transcription polymerase chain reaction (RT-PCR) on nasopharynx swab for 16 patients, confirming COVID-19 in eleven cases, and serological test in 3 cases, all of which were positive. For one subject, SARS-CoV-2 infection was confirmed by immunohistochemistry on a specimen of the resected small bowel, while for two patients the disease was suspected based on clinical features and pulmonary findings at imaging, despite a negative RT-PCR on nasopharynx swab. For 5 patients, the diagnostic test used was not reported by the authors.
Diagnostic workup for inherited or acquired thrombophilia was reported in 11 cases, two of which had positive lupus anticoagulant (LAC) (one at low titer), and one an established diagnosis of essential thrombocythemia (ET). Further tests performed were reported as negative by the authors.

Treatment and outcomes
Before SVT was diagnosed, three patients were taking anticoagulation therapy (one at therapeutic dose). After the diagnosis, an anticoagulation treatment was started in 16 cases, including unfractionated heparin (UFH) or low molecular weight heparin (LMWH) (11 patients), and Apixaban (2 patients). In three case, the anticoagulant treatment introduced was not specified. Seven patients with radiological features of intestinal ischemia underwent bowel resection, one of whom had also thrombolysis and thrombectomy in the context of upper mesenteric artery thrombosis.
Outcome was available only for 14 patients, four of whom died. These were 2 males and 2 females, aged 79, 62, 61, and 42, respectively. One subject had no known comorbidities, another one had only diabetes reported, the third one had diabetes, obesity, arterial hypertension and cirrhosis, while the last one had extreme obesity and a ventriculoperitoneal shunt due to a partially resected craniopharyngioma. All these patients had multivessel involvement, including upper mesenteric artery, as well as bowel ischemia complicating SVT.
Two subjects had clinical worsening following anticoagulant therapy (one due to subsequent bowel ischemia, and the other one due to tight stenosis of mid-jejunum caused by congenital adhesion band). Both underwent bowel resection and were discharged thereafter. In one case, a 6-week follow-up imaging showed an established PVT with collateralization extending into the upper abdomen. At last, two patients had gastrointestinal bleeding after anticoagulant treatment introduction.

Discussion
Venous thromboembolism is a common complication of inpatients with COVID-19, with a prevalence as high as 21% in a recent meta-analysis of over 8 thousand patients [28] . The mechanisms underpinning the strong relationship between SARS-CoV-2 infection and venous and arterial thromboembolism are not clear yet, and likely include endothelial dysfunction, excessive inflammatory response, and hemodynamic components (stasis) [29 -33] .
There is no specific data regarding SVT, but it can be hypothesized that the above-mentioned mechanisms in patients with predisposing conditions such as chronic liver disease or MPD can lead to a preferential splanchnic location of thrombosis ( Fig. 1 ). Interestingly, however, in a meta-analysis by Diaz et al. on histopathological reports from deceased COVID-19 patients undergoing autopsy or liver biopsy, almost 30% of cases presented hepatic vascular thrombosis in spite of a low prevalence of known chronic liver disease [34] .
In the here reviewed cases, male gender was prevalent with a median age of 43 years. Current literature shows that the mean      age of patients developing venous thromboembolic events other than SVT during COVID-19 ranges between 60 and 70 years [35 , 36] . Concerning SVT in other clinical scenarios, the mean age of patients is estimated between 50 and 60 years in case of PVT [37] , and between 70 and 79 years for MVT [38] . Therefore, the data reported herein seem to suggest an uncommonly young age in patients developing SVT and concomitant COVID-19. Clinical presentation of SVT was often unspecific, especially in a context of multiorgan involvement and failure, making its prompt identification challenging. Typical presentation for patients without COVID-19 includes abdominal pain, reported in almost half of cases, gastrointestinal bleeding and ascites [39] . Abdominal pain was also the main reported symptom in the reported series, followed by vomiting and fever.
As for the known risk factors for SVT, only two patients had known liver cirrhosis. In one case, imaging tests performed at admission documented multifocal liver lesions consistent with hepatocellular carcinoma in the setting of newly discovered chronic hepatitis B. Advanced liver disease is itself a risk factor for SVT. Defining a clear etiological role of SARS-CoV-2 different in this population from SVT in non-cirrhotic patients remains unclear. However, it is likely that SARS-CoV-2 represents a further trigger in this context as well. Indeed, a worsening effect of COVID-19 on the prognosis of patients with cirrhosis is well established and progression of PVT in patients despite prophylactic therapy with LMWH has been reported [40] .
Interestingly, in one case COVID-19 diagnosis was possible with RNAs in situ hybridization technique applied to the resected small bowel, suggesting that local inflammation due to SARS-CoV-2 infection might be a major trigger in the development of SVT for these patients.
The diagnostic workup for inherited or acquired thrombophilia led in one case to a new diagnosis of ET, while positive LACs were found in two patients. In non-cirrhotic SVT, the mean prevalence of JAK2 V617F mutation ranges between 41.1 and 27.7%, while the mean prevalence of MPN is estimated at 40.9% and 31.5% for BCS and PVT, respectively. Additionally, for almost 20% of the subjects with SVT, JAK2 V617F screening identified MPN in patients otherwise with no typical feature of MPN [41] .
With regard to LAC positivity, recent observations have suggested that single LAC positivity is a common finding during the acute phase of SARS-CoV-2 infection, without a clear causal relationship with thrombotic events. Conversely, other high-risk thrombophilia conditions, such as triple antiphospholipid antibodies positivity or high anticardiolipin/antibeta2-glycoprotein I antibodies, have rarely been described in this setting [42] . Importantly, since these findings have often been transient, unconfirmed at later measurements and not consistently associated with throm- botic events, they may not be an adequate screening tool for acquired thrombophilia in the acute phase of the disease [43] .
In the here reviewed cases, SVT was assessed using DUS imaging only for one patient, whereas in the rest of the cases CT scans ruled out the diagnosis. One patient, a young woman who was later on diagnosed of ET, showed signs of cavernous transformation of the portal vein and large porto-systemic collaterals, suggesting the presence of an acute on chronic thrombosis. DUS is a validated technique in detecting SVT with an accuracy up to 90% for diagnosing PVT, cavernous transformation of the PV and BCS [44] , although the sensitivity of DUS is lower for mesenteric and splenic veins [45] . CT scan or similar cross-sectional imaging should be always considered as diagnostic tool, which may also be used to investigate the presence of signs of bowel ischemia [46] and assess the extent of involvement of all the vessels of the PV system, since involvement of more than one vessel indicates a worse prognosis [47 , 48] . In 8 of the SVT here reviewed cases, thrombosis involved more than one vessel.
Of note, three patients were on anticoagulation at SVT diagnosis, of whom one at therapeutic dose for concomitant deep venous thrombosis (unclear if in the outpatient or inpatient setting). One patient was receiving LMWHs at standard prophylactic dose in a COVID-19 Unit, and another one in the outpatient setting. In the only reported case of SVT occurring during ICU stay, the presence of pharmacological prophylaxis of thrombosis was not reported. Previous research suggested that the incidence of thrombotic complications in ICU patients with COVID-19 infections is remarkably high despite anticoagulant therapy at standard prophylactic dose [49] . Notwithstanding this, recent evidence on patients admitted to the ICU with COVID-19 showed that intermediate-dose prophylactic anticoagulation did not result in a significant difference in the primary outcome of a composite of adjudicated venous or arterial thrombosis, treatment with extracorporeal membrane oxygenation, or mortality within 30 days [50] . This might also hold true as for SVT. Ongoing trials should clarify the role of different prophylaxis strategies in the outpatient setting, including LMWHs at standard prophylactic dose, DOACs at both low and high intensity, aspirin, and sulodexide [51] .
The goal of treatment of acute SVT is to achieve the patency of the vein, thus preventing bowel infarction, liver injury, and late complications of portal hypertension. The timing for starting an anticoagulant therapy is crucial in order to avoid potentially lifethreatening gastrointestinal bleeding. All but five patients included in the present systematic review received anticoagulant therapy. However, four patients who underwent urgent intestinal resection have died, thus supporting the importance of immediate surgical evaluation in subjects with severe abdominal clinical presentation (e.g. bowel infarction at imaging, peritonitis, septic shock) even before considering the anticoagulant treatment. Preventive evaluation of signs of portal hypertension including gastroesophageal varices should be evaluated in a case-by-case basis in this context [46] .
Evidence on which anticoagulant therapy should be used in patients with SVT is limited and choice is based mostly on clinical experience. UFH, LMWH and vitamin K antagonists are commonly used. The use of direct oral anticoagulants (DOACs) still remains off-label in most countries for SVT [46] . Potential malabsorption in case of intestinal ischemia should be always considered as a potential risk of lack of efficacy for oral therapy. Current recommendations for in-hospital patients with COVID-19 requiring anticoagulation suggest LMWH as first-line treatment [52] , emphasizing its higher stability compared with UFH, particularly during the cytokine storm phase, and its reduced risk of interaction with antiviral drugs compared with DOACs. Indeed, antiviral therapy for COVID-19 has been reported to dramatically increase DOACs plasma levels [53] . Accordingly, the use of LMWH may constitute the most effective and safe strategy also for patients with SVT during COVID-19. Future research is needed to clarify these aspects.
Limitations of the present systematic review include the small number of cases reported, and reporting bias (likely reporting of the most severe cases). The conclusions of the present study rely on the quality and accuracy of the reports included in the analysis.

Conclusions
SVT has been reported in 21 COVID-19 cases so far, and as such it can be considered as an uncommon manifestation of SARS-CoV-2 infection. However, SVT is often fatal, thereby requiring prompt recognition and treatment. Young patients and subjects without known comorbidities may be at risk of developing this complication. A high level of warning should be raised in presence of SVTcompatible symptoms in the setting of COVID-19.
Particular attention should be given to screening of inherited or acquired thrombophilia, bearing in mind the correct timing for testing. DUS and cross-sectional imaging remains essential for diagnosis and mapping of thrombosis extent in the portal venous system. Careful monitoring of potential signs of bowel ischemia should be performed, in order to provide appropriate treatment at early stages.
Future perspectives might embrace long-term follow-up of patients with SVT and COVID-19, aiming to evaluate its natural history, including development of late portal hypertension, specific histological alterations of the liver parenchyma, and vessels recanalization rates. In this sense, an international registry could be an extremely useful tool to document and group such sporadic cases, and follow-up the clinical evolution of these patients over time. Identifying the pathophysiological mechanisms underlying the relationship between SARS-CoV-2 infection and SVT may also be of great interest.

Declaration of Competing Interest
Giacomo Buso, Chiara Becchetti, Annalisa Berzigotti have no conflict of interest to declare.

Funding
Chiara Becchetti received financial support from the Stiftung für Leberkrankheiten.