Hematopoietic Stem Cell Transplantation in Patients With Myelofibrosis and Splanchnic Vein Thrombosis: A Case Series

Abstract

Myeloproliferative neoplasms (MPN) are associated with increased risk of splanchnic vein thrombosis (SVT), which contributes to morbidity and mortality. SVT, including portal, superior mesenteric, splenic, and hepatic vein thrombosis, disrupts portal pressures, leading to the development of portal hypertension (PHT) and complications such as varices, splenomegaly and liver failure, which increase the morbidity and mortality. The impact of prior SVT on outcomes of patients undergoing allogeneic hematopoietic stem cell transplantation (HCT) for myelofibrosis (MF) is not well understood. Although a small case series found a strong correlation between SVT and hyperbilirubinemia [1], the data is limited, making the optimal management of anticoagulation, hepatic comorbidities, and PHT sequelae challenging in this patient population. This study evaluates the impact of pre-existing SVT and its treatment on transplant outcomes.

We screened 334 consecutive patients who underwent HCT for MF at The University of Texas MD Anderson Cancer Center between January 2000 and August 2023 and identified 12 with pre-existing SVT. Patient characteristics, liver and thrombosis-related data, transplant details and outcomes were collected through retrospective review. Model for End-Stage Liver Disease (MELD) scores were calculated using creatinine, bilirubin, INR, and sodium levels to stratify the degree of chronic liver dysfunction. Continuous variables are reported as median and range, while categorical variables as number and percentage. Overall survival was estimated using the Kaplan–Meier method.

With regards to patient characteristics, the median age was 53 years (range: 39–73). Five patients had primary MF, four had post-essential thrombocythemia (PET-MF), and three had post-polycythaemia vera (PPV-MF). The majority (n = 8, 66.6%) were JAK2 V617F-positive. One patient was MPL positive. Patients received conditioning regimens based on institutional protocols, considering age and comorbidities. Individual characteristics and outcomes for all patients are summarized in Table 1. Other baseline characteristics are seen in Table S1.

TABLE 1. Detailed patient characteristics.

Patient	Age at transplant	Sex	JAK2 status	Additional mutations via NGS	Thrombosis type	Days SVT prior to transplant	Splenectomy prior to thrombosis	Yerdel classification	Anticoagulation	Date of transplant	Conditioning regimen
1	40	F	Positive	No	Splenic vein thrombosis	109	No	NE	No	3-Jul-23	Bu/Cy
2	70	F	Positive	No	Portal vein, SMV and splenic vein thrombosis	145	Yes	3	Yes, 120 days	14-Sep-22	Flu/Mel
3	72	M	Positive	ASXL1, EZH2, GATA2, SF3B1	Portal vein thrombosis	74	Yes	1	No	29-Aug-22	Flu/Bu
4	52	F	Positive	TP53	Portal and SMV thrombosis	503	Yes	2	Yes, 180 days	11-Sep-19	Flu/Bu
5	51	F	Negative	MPL	Portal vein thrombosis	234	No	2	No	9-Sep-19	Flu/Bu
6	39	F	Positive	No	Portal vein thrombosis, splenic vein thrombus	613	No	3	Yes, 2 years	30-Aug-16	Flu/Mel/TBI
7	46	M	Positive	NE	Portal vein, SMV and IVC thrombosis	4018	No	3	No	13-Dec-13	Flu/Mel
8	54	F	Positive	NE	Portal vein thrombosis	4748	No	3	No, portocaval shunt	15-Dec-09	Flu/Bu/Thiotepa
9	54	F	Positive	NE	Portal and splenic vein thrombosis	41	Yes	NE	Yes, 252 days	14-Sep-06	Flu/Bu/Thiotepa
10	73	M	NE	NE	Portal vein thrombosis	46	Yes	3	Yes, 95 days	24-Oct-05	Flu/Bu/Thiotepa
11	62	M	NE	NE	Portal and splenic vein thrombosis	197	Yes	1	Yes, 125 days and TIPS	9-Apr-03	Flu/Bu/Thiotepa
12	49	F	NE	NE	SMV and splenic vein thrombosis, acute hepatic vein thrombus post-HCT	1397	No	NE	Yes	12-Dec-00	Flu/Bu/Thiotepa
(Continues)

Patient	Donor type	MELD score	Portal hypertension	Ascites	Varices	Variceal bleeding pre-transplantation ± intervention	Beta blocker use	GI bleeding post transplantation	Liver outcome in first 100 days	Clinical course
1	Haplo	8	Yes	Yes	Esophageal	No	Yes	Major GI bleed day 3, 4 and 11, requiring splenic artery embolization	NAD	Uncomplicated
2	MUD	8	Yes	No	Esophageal	No	No	No	NAD	Uncomplicated
3	MSD	7	Yes	Yes	Esophageal, gallbladder	No	No	No	NAD	Uncomplicated
4	MUD	8	No	No	No	No	No	No	NAD	Uncomplicated
5	Haplo	15	Yes	Yes	Esophageal, gastric	Hematemesis requiring multiple banding procedures	Yes	No	Severe hyperbilirubinemia and VOD (day 4)	Decompensated liver failure with resultant encephalopathy (VOD), platelet refractoriness and intracranial hemorrhage
6	MUD	9	Yes	Yes	Esophageal, gastric	No, pre-emptive banding	Yes	Minor GI bleed day 1386	NAD	Increased esophageal, splenic and gastric varices and splenomegaly
7	MMUD	17	Yes	Multiple ascitic paracenteses	Esophageal	Yes, variceal banding	Yes	Recurrent major GI bleeding days 12, 68, 84, 105, 111	Grade I hyperbilirubinemia, Moderate transaminitis	Pleural effusions requiring recurrent drainage, progressive renal failure and death due to fungal pneumonia
8	MSD	13	Yes	Yes, requiring paracentesis	Esophageal	Yes	Yes	Recurrent major GI bleeding day 14 and 3223	Grade I hyperbilirubinemia	Decompensated cirrhosis requiring paracentesis, liver GVHD. Relapse day 254
9	MSD	9	No	No	No	No	No	No	NAD	Lost to follow up 2009
10	MSD	11	Yes	Yes	No	Yes	No	No	Grade I transaminitis	Died of secondary malignancy
11	MSD	10	Yes	Yes	No	Yes, variceal banding	No	No	NAD	Recurrent admissions with encephalopathy and ascites requiring TIPS revisions
12	MSD	13	Yes	Yes, requiring paracentesis	Esophageal, gastric	No	No	Yes, day 236	NAD	Recurrent admissions with ascites and encephalopathy requiring TIPS, developed severe GI bleeding with coagulopathy, liver failure

Half of the patients developed SVT following splenectomy, while the other half developed de novo SVT. Three developed SVT within 100 days pre-HCT, all post-splenectomy. The median time from thrombosis to transplant was 215.5 (range: 41–4748) days. The median Yerdel score, describing the extent of thrombosis and risk of complication, was 3 (range: 1–3). Six (50%) patients had chronic SVT and were not anticoagulated pre-HCT. The remaining six (50%) received anticoagulation (DOAC or enoxaparin) for acute thrombosis, with a median duration of 180 days (95–1950). Three (42.8%) continued anticoagulation peri-HCT after platelet recovery. Further information is seen in Table S2.

With regards to liver-related characteristics, the median MELD score was 9.5 (range: 8–17). Most patients had radiologic evidence of PHT (n = 10; 83.3%) or varices (n = 8; 66.6%). Of those with varices, six (75%) had pre-HCT variceal bleeding. Two had pre-emptive banding of varices, and four received beta blocker prophylaxis. Two patients required pre-HCT intervention with transjugular intrahepatic portosystemic shunt (TIPS) and portocaval shunt. No biopsy-proven cirrhosis or iron overload was found pre-HCT, though biopsies were only performed in 4/12 patients. No alternate causes of liver dysfunction were identified. Further information is seen in Table S3.

At Day 100, 11 (92%) patients were alive and disease-free. At a median follow up of 47.5 months, seven (66.6%) were alive and disease-free. Five-year overall survival was 64% (SE 14%). Five patients died: one from secondary malignancy (Day 2315), one from fungal infection related to poor graft function (Day 148), although recurrent gastrointestinal (GI) bleeding and ascites contributed to their decline, and three from direct liver-related complications (veno-occlusive disease [VOD] and intracranial hemorrhage, Day 12; liver failure on Day 103 and liver failure on 260).

For those patients who died from direct liver-related complications: the first patient (MELD 15, pre-HCT variceal bleeding requiring banding) developed severe VOD at Day 4 post-HCT and died of intracranial hemorrhage at Day 12. The second patient (MELD 10, pre-HCT variceal bleeding requiring banding) suffered liver failure following TIPS placement for portal and splenic vein thrombus at Day 62, leading to death on Day 103. The third patient (MELD 13) developed symptomatic ascites post-HCT with imaging revealing acute hepatic vein thrombus (BCS) and cavernous portal vein thrombus requiring TIPS placement on Day 95. She developed progressive liver failure and GI bleeding, dying at Day 260. Overall, liver-related complications were associated with pre-HCT MELD scores > 10, pre- or post-HCT variceal bleeding, TIPS placement, and failure of recanalization on post-HCT imaging.

Non-fatal liver-related complications included liver toxicity and progressive thrombosis. Grade 1 hyperbilirubinemia (n = 3; 25%) and Grade 1 transaminitis (n = 2; 17%) resolved without intervention by Day 100. Five patients (42%) developed long-term sequelae of PHT at a median of 146.5 days (range: 72–3304). These included decompensated ascites (n = 3; 33.3%), variceal bleeding requiring endoscopic management (n = 4; 33.3%), and encephalopathy in the setting of previous TIPS (n = 2; 16%). Median spleen size was 24 cm, and 75% had not undergone splenectomy. Recurrent GI bleeding post-HCT (median 76 days, range: 3–3323) occurred in patients with pre-HCT esophageal varices and no anticoagulation, despite beta blocker prophylaxis. New venous thromboembolism (VTE) was noted in four patients post-HCT, including three (75%) occurring while off anticoagulation. Three (75%) of these events occurred outside the splanchnic system (deep veins of the limbs and pulmonary embolism). Recanalization of SVT occurred in only 16% of patients.

Despite this, some patients had encouraging clinical outcomes at extended follow up. Patient 8 (MELD 8) experienced recurrent variceal bleeding and ascites requiring paracentesis up to 9 years post-HCT, but was alive and clinically stable at 12.5 years post-HCT with long-term hepatology follow-up and diuretic therapy. Patient 6 (MELD 9) developed increasing varices and splenomegaly post-HCT but remained disease-free with no liver toxicity 2577 days post-HCT.

This case series indicates that MF patients with pre-existing SVT can successfully undergo HCT, although they face liver and thrombosis-associated morbidity and mortality. Optimal management of SVT, particularly in the peri-HCT period, is not well studied. However, the rate of recurrent thrombosis in this setting is increased [2]. While anticoagulation in acute thrombosis improves recanalization and reduces PHT, it may increase bleeding, particularly in the setting of varices. Consideration of beta blockade or preemptive banding may mitigate this risk, although this has not been validated in MF. Chronic SVT rarely recanalizes, so life-long anticoagulation is generally recommended to reduce risk of recurrence. In our cohort, long-term anticoagulation, including peri-HCT, was not associated with bleeding, though the decision to continue anticoagulation peri-HCT should be individualized. Results with interventions such as TIPS, thrombolysis, and liver transplantation are variable in non-HCT patients. Portosystemic shunting shows a survival benefit in BCS patients, but its benefits in SVT remain unclear [3].

Management of splenomegaly pre- and peri-HCT is challenging. Splenectomy is often used in an attempt to reverse PHT and improve engraftment, however post-operative SVT rates are as high as 55% with laparoscopic approaches [4]. Morbidity may be significant, including intestinal ischemia and even death. Risk factors for SVT development include splenic weight, co-existent MPN, splenic/portal vein diameter, low white blood counts, and anatomic variation [5]. Half of the patients in our cohort developed SVT following splenectomy, but splenectomy itself did not appear to be associated with long-term morbidity.

MF patients are at risk of post-HCT hepatotoxicity due to iron overload, extramedullary hematopoiesis, and pre-existing PHT. Wong et al. found an increased risk of VOD in patients undergoing HCT for MF, with an association between SVT and hyperbilirubinemia [1]. In contrast, our data showed no persistent liver dysfunction after Day 100, suggesting that with appropriate patient selection, hepatic function is not compromised. One patient died due to complications of VOD, indicating the need for careful surveillance in the post-HCT period.

PHT is a common complication of MF, with potential etiologies including splenomegaly, PVT, extramedullary hematopoiesis and recently, sinusoidal fibrosis [1]. Resultant varices increase the risk of hemorrhage and ascites, leading to long-term morbidity. Thrombosis often contributes, with nearly half of non-HCT patients with JAK2 V617F-associated non-cirrhotic PVT developing PHT during follow-up [6]. In our cohort, sequelae of PHT were evident at Day 100, with persistent morbidity evident even after 12 years in some patients. Liver-related variables such as bilirubin, albumin, and INR, MELD score (> 10) and non-recanalization pre-HCT were key prognostic indicators in this cohort. Other high-risk clinical characteristics, including pre-HCT variceal bleeding, appear to predict similar bleeding post-HCT, regardless of intervention. Pre-HCT optimization, such as variceal banding, may improve outcomes, although this is unclear. All patients undergoing HCT workup should have dedicated imaging for PHT, endoscopic evaluation for variceal treatment and long-term multidisciplinary follow up.

Our study, although limited by its small patient cohort, retrospective design, and the lack of representation of BCS or hepatic vein thrombus, suggests that HCT is feasible and curative for selected patients with SVT with low early and overall mortality. However, the long-term sequelae of PHT remain a concern. Factors such as variceal bleeding and MELD score > 10 appear to increase the risk of future morbidity and mortality related to PHT. Multidisciplinary pre-HCT evaluation including dedicated imaging for PHT, screening and pre-emptive treatment of varices, and long-term follow up contribute to successful outcomes in this patient population. Larger studies are needed to validate these results and refine risk stratification for patients with SVT undergoing HCT.