P1020: SCREENING FOR PORTAL HYPERTENSION IN PATIENTS WITH PHILADELPHIA NEGATIVE MYELOPROLIFERATIVE NEOPLASMS

Abstract

Background: Portal hypertension (PH) is a well-recognized complication of BCR-ABL negative myeloproliferative neoplasms (MPNs) and is associated with significant morbidity and mortality. However, the prevalence of PH and its associated risk factors have not been studied systematically in this patient population. Screening by esophagogastroduodenoscopy (EGD) can facilitate early diagnosis of PH and opportunity for intervention to prevent complications. Aims: To investigate the prevalence of PH and its patient and disease-related risk factors. Methods: From May 2013 to March 2019, MPN patients ≥ 18 years old with spleen >5 cm below the costal margin in myelofibrosis (MF) or any size in Polycythemia Vera (PV) or Essential Thrombocythemia (ET) were enrolled prospectively (NCT01816256). Patients with a known history of portal or hepatic vein thrombosis, liver cirrhosis, and esophageal varices were excluded. Screening for PH was done by EGD to evaluate for esophageal varices (EV) or gastric varices (GV), and portal hypertensive gastropathy (PHG). Abdominal Doppler ultrasonography (USG) was performed to evaluate for portal vein thrombosis and spleen size. PH severity was based on the size of varices or severity of PHG. For sample-size calculation, a precision approach was used, considering scenarios where PH prevalence ranged from 5 to 10% and an upper limit for sample size of 100. At a prevalence estimate of 8%, the 90% confidence interval (CI) was 4-14%, with an acceptable lower limit for moving forward with a screening program. Prevalence of PH was calculated as a proportion and CI are exact using Clopper-Pearson method. Results: Among 102 enrolled patients, EGD and USG were completed in 85 and 84, respectively. Primary and Post-PV/ET MF comprised the majority of the study population (93% [79/85]). Evidence of PH was found in 32 patients (38%, [90% CI 29-47]) of whom 12 (14% [90% CI 9-22]) had moderate to severe PH. EV, GV, and PHG were found in 81% (26/32), 13% (4/32), and 31% (10/32) of patients, respectively. The prevalence of PH among MF and PV/ET patients was 35% (28/79) and 67% (4/6), respectively. In MF patients, PH was moderate-to-severe in 13% (90% CI 8-20). No cases of portal vein thrombosis were identified on USG. Apart from a palpable liver observed more frequently in non-PH patients, there were no other significant differences in baseline characteristics among patients with and without PH. The frequency of JAK2, MPL and CALR mutations among PH patients was 78% (25/32),13% (4/32), and 6% (2/32). There was a trend towards more JAK2 mutations in PH patients (78% [25/32] vs 64% [34/53], p=0.23), but no difference in DIPSS risk or frequency of high molecular-risk mutations. Overall survival was similar in both groups. PH-directed interventions were initiated in 59% (19/32); 28% (9/32) were started on beta blocker (BB) prophylaxis, 16% (5/32) were treated with BB and variceal banding, and 16% (5/32) continued on surveillance EGDs alone. Three patients with moderate PH at baseline developed variceal bleeding (2) and ascites (1). All 3 patients were on BB prophylaxis, and 1 had prior prophylactic banding. There were no adverse events related to study procedures EGD or USG. Image:Summary/Conclusion: Given the study population, our results are generalizable to MF patients. Our study shows that clinically significant PH is common in MF patients, and led to additional clinical interventions in significant number of patients. These data support the inclusion of asymptomatic PH screening in the work up of MF patients.