BACKGROUND:Gastrointestinal hemorrhage (GIH) is reported to occur in 1-8% of patients admitted with acute ischemic stroke (AIS). AIS is considered to be a relative contraindication to GIE. AIMS:Evaluate the outcomes of gastrointestinal endoscopy (GIE) in patients hospitalized with AIS and GIH. METHODS:Patients hospitalized with AIS and GIH were included from the National Inpatient Sample 2005-2014. Primary outcome measure was in-hospital mortality in patients with AIS and GIH who underwent gastrointestinal endoscopy. Secondary outcomes were (1) resource utilization as measured by length of stay (LOS) and total hospitalization costs and (2) to identify independent predictors of undergoing GIE in patients with AIS and GIH. Confounders were adjusted for by using multivariable regression analysis. RESULTS:A total of 75,756 hospitalizations were included in the analysis. Using a multivariate analysis, the in-hospital mortality was significantly lower in patients who underwent GIE as compared to those who did not [aOR: 0.4, P < 0.001]. Patients who underwent GIE also had significantly shorter adjusted mean LOS [adjusted mean difference in LOS: 0.587 days, P < 0.001]. Patients with AIS and GIH who did not undergo GIE had significantly higher adjusted total hospitalization costs. [Mean adjusted difference in total hospitalization costs was $5801 (P < 0.001).] Independent predictors of undergoing GIE in this population were male gender, age > 65 years, Asian or Pacific race, hypovolemic shock, need for blood transfusion and admission to urban non-teaching hospital. CONCLUSIONS:Gastrointestinal endoscopy can be safely performed in a substantial number of patients with AIS and GIH.

Introduction: Higher volume of endoscopic procedures by a provider may impact quality of colonoscopy. Adenoma detection rate (ADR) among primarily endoscopists has been reportedly 10% higher than of non-endoscopists. We performed a retrospective, single institution study to assess differences in colonoscopy quality indicators between gastroenterologists and transplant hepatologists. Methods: Data was collected from colonoscopies performed at Westchester Medical Center between Jan 1, 2017 and Dec 31, 2017. Patients with documented familial adenomatous polyposis, Lynch syndrome, BRCA1 gene mutation, inflammatory bowel disease or gastrointestinal bleeding were excluded. We studied patient demographics age, gender and race, indication for colonoscopy, particulars of the performing physician and adequacy of bowel preparation were also documented. Results: We collected data from 623 colonoscopies performed by 6 gastroenterologists (N=559) and 2 transplant hepatologists (N=64). 48.6% (N= 303) were performed for average risk screening and 31.9% (N=199) were performed due to prior history of colon polyps<./p>Mean number of years since graduation from GI fellowship was 17.5 years for hepatologists and 21.5 years for primary gastroenterologists. Likelihood of adenoma removal was higher for gastroenterologists (35.2%) than hepatologists (31.3%, Table 1). For screening colonoscopies, the difference in ADR between gastroenterologists (33.6%) and hepatologists (18.8%) was more pronounced but not statistically significant. Unadjusted odds ratio for adenoma removal among gastroenterologists was 2.192 for screening colonoscopies and 1.197 for all colonoscopies. Gastroenterologists (all colonoscopies: 89.8%, screening colonoscopies: 88.2%) were significantly more likely to have adequate bowel preparation than hepatologists (all colonoscopies: 73.4%, screening colonoscopies: 71.9%). Conclusion: Gastroenterologists had a trend toward higher ADR and significantly higher likelihood of adequate bowel preparation compared to transplant hepatologists. It is possible that hepatologists' patients were sicker, leading to a desire to minimize sedation time and defer removal of small adenomas due to underlying bleeding risk. Furthermore, there is data to suggest that patients with advanced liver disease are more likely to have inadequate bowel preparation. Hepatologists may wish to prescribe additional medication for bowel preparation and support staff can improve the quality of bowel prep by optimizing patient education.1120_A Figure 1. Comparing proportions of adenoma detection among attendings1120_B Figure 2. Comparison of proportions of adequate bowel prep among attendings1120_C Figure 3. Unadjusted odds ratio associated with adenoma detection for physician characteristic

Esophageal varices can cause life-threatening complications and are most often a sequela of liver disease. Although a rare cause of gastrointestinal bleeding, downhill variceal bleeding secondary to superior vena cava (SVC) obstruction should be considered in the differential diagnosis for patients with upper gastrointestinal hemorrhage. We discuss two such cases of downhill esophageal varices presenting with hematemesis in patients with end stage renal disease and no history of cirrhosis. These varices were thought to be secondary to SVC occlusion caused by complications from previous dialysis catheters. However, their difficult anatomy posed a significant challenge to the therapeutic interventions.

Gastrointestinal (GI) hemorrhage is a common complication in patients with cerebrovascular accidents. Reported incidence has ranged from 1% to 8% in several published studies. We studied a national inpatient database to assess the effects of GI endoscopy on mortality and length of hospital stay among patients with acute stroke.

Introduction: Prior studies that evaluated the seasonal variation of upper gastrointestinal bleeding have led to no clear consensus. Some reports demonstrated seasonal variation, possibly with seasonal fluctuations in NSAID use, while other studies do not show a seasonal relationship. We reviewed the NIS database to assess seasonal differences in hospitalizations due to variceal bleeding. Methods: Method: We conducted a cross-sectional study using NIS data for years 2005 to 2014. We used single level CCS diagnosis code(153) to identify cases with GI hemorrhage and then used ICD-9-CM code (456.0 & 456.20: esophageal varices with bleeding) to identify cases with a primary or secondary diagnosis of variceal bleeding. We performed Chi-square analysis to study the demographics, and a multivariate to compare in-hospital mortality. Adjustments were made for the confoundings: age, gender, race, comorbidities, LOS, admission status and hospital factors including location and teaching status. All statistical analysis was performed with SPSS v25.0 (IBM Corp, Chicago, IL) Results: A total of 348,958 hospitalizations with variceal bleeding were reported between 2005 and 2014. The mean age of the entire cohort was 55.79 (SD:12.15), 69.1% were male. The mean LOS for the entire cohort was 6.35 days.The highest number of variceal bleed-related hospitalizations were reported in December (N=30,786; 8.8%) followed by March (N=30,492; 8.7%) and January (N=30,404; 8.7%). The lowest number of hospitalizations were reported in June (27,247; 7.8%). We calculated adjusted odds ratio for in-hospital mortality associated with months. Mortality was highest in January (Adjusted OR=1.268, 95% CI:1.194-1.345) The greatest number of esophagogastroduodenoscopy (EGD) for variceal bleeding were in December (N=25,940), out of which 77.5% (N= 20,129) involved endoscopic therapy for hemoastasis. The fewest EGD procedures were performed in June (N=22,849), out of which 75.1% (N=17,176) involved endoscopic therapy. In-hospital mortality for the patients undergoing EGD was 8.2%, while mortality was 21.5% for patients who did not undergo EGD. There was no significant difference in LOS or cost across seasons. Conclusion: There appears to be a seasonal variation in the incidence of variceal bleeding in the United States. The reasons for these variations remains unclear. This study is limited in that the analysis is based on the accuracy of diagnostic codes reported at the time of hospitalization.358_A Figure 1. Overall Monthly Incidence and Mortality of Variceal- Bleeding from 2005 to 2014358_B Figure 2. Graphical presentation of variation of in-hospital mortality with Esophageal variceal bleeding.358_C Figure 3. Multivariable regression analysis and adjusted Odds ratio for in-hospital mortality associated with each month.

Introduction: It has been reported in the literature that gastroenterology fellow involvement during colonoscopy improves the adenoma detection rate (ADR), thus improving quality of colon cancer screening. [2, 3] We performed a review of colonoscopy procedures at a single teaching hospital to compare fellow adenoma removal percentages with those reported in the literature. [1, 3] Methods: We performed a retrospective chart analysis of the electronic medical record and collected data from colonoscopies performed between January 1, 2017 and December 31, 2017. Patients with documented familial adenomatous polyposis, Lynch syndrome, BRCA1 gene mutation, inflammatory bowel disease or gastrointestinal bleeding were excluded. We studied patient demographics, age, gender, race, indication for colonoscopy, particulars of the performing attending, fellow participation and adequacy of bowel prep were also documented. Results: We collected data from 623 colonoscopies performed at our institution. Out of total 623 colonoscopies, 48.6% (N= 303) were performed for average risk screening and 31.9% (N=199) were performed due to history of colon polyps. The next most common indications were family history of colon cancer (N=55) and personal history of colon cancer (N=21). Mean age of the cohort was 60.6 years. Approximately half (N=313) of the patients were female. The cohort was 60.5% white, 17% black, 15.9% Hispanic and 4% Asian. Fellows were involved in 54.4% (N=176) of colonoscopies included in this study. For screening colonoscopies, ADR for colonoscopies without a fellow was 26%, which improved to 30.8% with first year fellows, 32.7% with second year fellows, 40.0% with third year fellows (Table 1). For 6 out of 8 attendings, ADR improved with involvement of a fellow.Unadjusted odds ratio for adenoma detection during screening colonoscopy was highest when the attending was accompanied by a third year fellow. The odds ratio was lowest when attendings performed colonoscopy without any fellows on the case. Conclusion: Involvement of a gastroenterology fellow in screening colonoscopies significantly increased ADR, hence improving overall quality of colorectal cancer screening. Further studies are needed to assess other outcomes associated with fellow involvement, such as withdrawal time and duration of anesthesia administration.492_A Figure 1. Proportion of adenoma removal by fellow training year.492_B Figure 2. Unadjusted Odds Ratio associated with adenoma detection rate for the fellow involvement.492_C Figure 3. Comparing adenoma detection rate with and without fellow for each attending.

Introduction: Adenoma detection rate (ADR) has been widely accepted by gastroenterologists as a surrogate measure of colon cancer screening quality of care. Previous studies have shown a strong inverse correlation between ADR and risk of colorectal cancer. We performed a single teaching institution analysis to assess differences in patient and physician-related factors and their association with ADR. Methods: We performed a retrospective chart analysis, collecting data from colonoscopies performed at Westchester Medical Center between January 1, 2017 and December 31, 2017. Patients with documented familial adenomatous polyposis, Lynch syndrome, BRCA1 gene mutation, inflammatory bowel disease or gastrointestinal bleeding were excluded. We studied patient demographics, age, gender and race, indication for colonoscopy, particulars of the performing attending and adequacy of bowel preparation. Results: We collected data from 623 colonoscopies performed by 6 gastroenterologists (N=559, 89.8%) and 2 transplant hepatologists (N=64, 10.2%). 48.6% (N= 303) were performed for average risk screening and 31.9% (N=199) were performed due to prior history of colon polyps. Mean age of the cohort was 60.6 yrs and half (N=313) of the patients were female. Out of 623 procedures, one or more adenomas were removed in 34.8% (N=217). Among 303 screening colonoscopies the ADR was 32% (N=97). Overall 88.1% of colonoscopies had adequate bowel prep. When comparing different attendings, ADR for the two hepatologists was 21.4% and 16.7%. Men were significantly more likely to have adenomas removed compared with women (40.3% vs 29.4%, p=0.006). There was a strong positive correlation between years of attending experience (calculated number of years since graduation from gastroenterology fellowship) and ADR with Pearson correlation value of 0.849 (p=0.008). Graphical presentation has been shown with a scatterplot Image 1. We performed a multivariable regression analysis and adjusted for age, gender, race, fellow involvement on the case and adequacy of bowel preparation (Table 1). Adequate bowel prep was the strongest factor associated with adenoma detection. Conclusion: ADR has a positive correlation with gastroenterologist years of experience. Gastroenterologists outperformed hepatologists in ADR and adequacy of bowel prep. Improving the quality of bowel prep would most likely increase ADR and boost quality of colon cancer screening249_A Figure 1. Scatter plot presenting relation between years of attending experience and adenoma detection rate249_B Figure 2. Multivariable regression analysis of clinical factors associated adenoma removal.

Aromatic antiepilectic drugs (AEDs) such as phenytoin, phenobarbital and valproic acid are associated with drug-induced hypersensitivity reactions. The hypersensitivity reaction to phenytoin can range from morbilliform skin rash to bullae to Stevens-Johnson syndrome and Toxic Epidermal Necrolysis. Metallic taste, gingival hyperplasia and hepatic injury are the most common gastrointestinal adverse effects of phenytoin. We present a case of esophageal ulceration leading to occult upper gastrointestinal (GI) bleeding as a rare adverse drug reaction (ADR) to phenytoin. 83 year old male with past medical history of stroke and seizure disorder presented with bleeding from the oral cavity, worsening diffuse body rash, bullae and facial swelling over a period of one month. He presented to an outside institution a few months prior due to bullous skin eruptions. Skin biopsy at that time suggested bullous hypersensitivity reaction due to phenytoin. The drug was discontinued and patient was placed on courses of oral corticosteroids with poor response. On presentation to our hospital, he was noted to have an extensive rash involving 54% of his body surface area, including the oral cavity. During his hospitalization, he had an acute drop in hemoglobin level from 9.1 g/dL to 6.9 g/dL as well as a positive stool occult blood test. Upper endoscopy was deferred due to comfort-oriented goals of care and concern for risk of injury to the oropharynx given extensive oral ulcerations. Upper GI series with barium was performed, showing patulous esophagus with shaggy carpet-like appearance of the esophageal mucosa and mucosal irregularity, including a 1 cm ulcer in the mid portion of the esophagus. Infectious and autoimmune work up was negative. He was started on a proton pump inhibitor. His rash responded well to high dose intravenous steroids, and the patient was discharged to a rehabilitation facility on an oral steroid taper. Phenytoin is a widely used AED that causes dose dependent ADRs due to its narrow therapeutic index. Phenytoin toxicity is common in patients with serum phenytoin levels >20 μg/mL. A high index of suspicion for phenytoin induced hypersensitivity reaction should be maintained when patients receiving phenytoin exhibit signs of upper GI bleeding or dysphagia. In patients exhibiting ADRs, discontinuation of phenytoin is highly recommended. Systemic corticosteroids may be required for patients with severe hypersensitivity reactions.Figure: Mucosal ulceration presenting as a filling defect.Figure: Multiple large mucosal defects with shaggy carpet like appearance of the esophageal mucosa.

Introduction: Colonic eosinophilia with focal cryptitis is a rare form of gastrointestinal (GI) pathology often associated with parasitic infections. Here, we report the case of an otherwise asymptomatic individual with occasional constipation who was found to have colonic eosinophilia with focal cryptitis on screening colonoscopy, leading to diagnosis and treatment of chronic Strongyloides infection of the lower GI tract. Case description: A 51-year-old Colombian man with past medical history of hypertension treated with amlodipine was referred from his primary care clinic for screening colonoscopy. He had no active complaints except occasional constipation. Physical exam was unremarkable. Lab work was significant for peripheral eosinophilia, with differential reporting 14.8% eosinophils and total white blood cell count of 8.6 K/mm3. Colonoscopy showed multiple 1-2 mm variably raised white spots in the transverse colon, sigmoid colon and rectum (Figure 1). A single sessile polyp, measuring 12 mm, was removed at the hepatic flexure by snare cautery polypectomy. All biopsies were negative for parasites but showed fragments of colonic mucosa with moderate eosinophilia and focal eosinophilic cryptitis (Figure 2). Stool for ova and parasites tested negative. Comprehensive testing to rule out secondary causes of eosinophilic colitis and cryptitis revealed positive serum IgG antibodies against Strongyloides stercoralis. Considering test results and recent travel to South America, chronic infection with Strongyloides was suspected. Patient was given a two-day regimen of ivermectin, with resolution of peripheral eosinophilia on repeat blood count as well as improvement in his occasional constipation.Figure: Colonoscopy showing multiple white spots in the colon and rectum.Figure: Histopathology showing fragments of colonic mucosa with eosinophilia and focal eosinophilic cryptitis (green arrows).Discussion: Colonic eosinophilia and focal cryptitis—if not due to the rare primary eosinophilic colitis—is associated with parasitic infections, drug allergies, connective tissue disorders or chronic inflammatory conditions such as inflammatory bowel disease. Strongyloidiasis, when chronic, is generally asymptomatic as most cases present with limited or vague symptoms. Immunocompromised individuals may develop hyperinfection syndrome and dissemination, which are associated with a mortality rate of up to 90% if left untreated. Hence, it is important for clinicians to have a high index of suspicion for Strongyloides infection, particularly in those with recent travel history to an endemic region, peripheral eosinophilia and relevant endoscopic findings.

We present a case of colonic mucosa-associated lymphoid tissue (MALT) lymphoma in a 62-year-old woman diagnosed after a positive test for fecal occult blood.

Introduction: Due to high demand for organ donation for renal transplantation, expanding the pool of potential donors is of major importance. Because of an increasing number of treatment options for hepatitis C (HCV) infection, individuals with chronic HCV have been identified as a group that could be considered for organ donation. Risk of graft loss and liver-related mortality have been shown to increase in kidney transplant patients with active HCV, but the arrival of highly effective HCV treatments offers considerable hope. Previous studies have suggested similar short-term patient and graft survival among HCV-positive recipients who receive an HCV-positive kidney compared with an HCV-negative kidney. The purpose of this study was to evaluate the outcomes of patients at our institution with chronic HCV infection who received a kidney from hepatitis C antibody positive donors. Methods: Six adult patients with end stage renal disease and chronic HCV infection underwent renal transplantation at Westchester Medical Center from HCV positive donors from January 2016 to February 2017. Data on patient age, gender, HCV genotype, fibrosis stage, HCV treatment regimen, achievement of sustained virologic response (SVR) and graft survival were collected. Results: Five patients were male and one was female. Ages ranged from 54 to 72. Fibrosis stage data by liver biopsy METAVIR score or FibroTest® was available for three patients; one had F0-1, one had F1-2 and one had F2-3 disease. Four patients had genotype 1a infection, one had genotype 1b and one had genotype 4. None of the patients experienced a change in HCV genotype post-transplant. All patients were started on direct acting antiviral therapy after transplantation (five received sofosbuvir/ledipasvir and the patient with genotype 4 infection received sofosbuvir/velpatasvir). Four patients achieved SVR at 24 weeks, while the other two patients are too early for SVR data. All patients have functioning kidney allografts. Conclusion: This study shows that transplantation of HCV positive donor kidneys to HCV viremic recipients was well tolerated in a cohort of patients at our institution. Once transplanted, these patients have access to an increased number direct acting antivirals and can usually achieve SVR. Extending donor criteria to include hepatitis C antibody positive patients can decrease the wait time for individuals on the renal transplant list, and should be considered in selected patients after obtaining informed consent.

Hepatitis C virus (HCV) is known for its oncogenic potential and has been found to be associated with hepatocellular carcinoma (HCC) and non-Hodgkin lymphoma. It has also been postulated that HCV may play a role in the development of other extrahepatic solid tumors of other organs of the body since it has been isolated from the vessel wall, kidney, and oral mucosa. In this article, we have reviewed epidemiological studies that have been done to look into the relationship of HCV with nonliver solid cancers of the pancreas, thyroid, renal, oral cavity, breast, and lung and nonpancreatic gastrointestinal cancers. Based on this review, HCV might be associated with an increased risk of renal cell and lung cancers.

A 69 year-old man with no significant past medical history presented due to four week history of fevers to greater than 101°F. He was initially diagnosed with a urinary tract infection and received a course of trimethoprim-sulfamethoxazole, but fevers recurred after the antibiotic was discontinued. He underwent abdominal MRI, which showed multiple ill-defined lesions within the liver as well as left portal vein thrombosis. Lesions grew in size and fevers continued despite a two week course of ertapenem as an outpatient. Laboratory evaluation was notable for white blood cell count of 12.0, elevated ESR of 121 and elevated CRP of 15 with normal hepatic function panel. Biopsy of one of the lesions was performed, showing a neutrophil-rich inflammatory infiltrate within and around the intrahepatic bile ducts suggestive of ascending cholangitis. Pan-lobular and multi-lobar areas of inflammation were also present, suggestive of an active infectious process. Computed tomography of the abdomen and pelvis was performed on the same day, showing subacute sigmoid colon diverticulitis with a small, contained perforation. The patient was placed on ciprofloxacin and metronidazole, with resolution of fevers and improvement in size of liver lesions on follow-up MRI two months later. Pylephlebitis is defined as septic thrombophlebitis of the portal vein. This rare condition is believed to be secondary to inflammation or possibly an interaction between the infectious organism and the venous endothelium causing activation of the clotting cascade. Pylephlebitis has been reported in association with a variety of intra-abdominal conditions, including diverticulitis, appendicitis, inflammatory bowel disease and pancreatitis. Symptoms may include fever, chills, abdominal pain, jaundice, fatigue, weight loss, nausea and vomiting. The limited number of case studies available on pylephlebitis have identified that a significant proportion of patients do not demonstrate localizing symptoms that identify the primary source of infection. This case is unusual due to the presence of both pylephlebitis and ascending cholangitis--the latter confirmed with liver biopsy. Abnormal lab tests may include positive blood cultures as well as elevated white blood cell count, transaminases, bilirubin and alkaline phosphatase. Treatment of pylephlebitis requires a prolonged course of broad spectrum antibiotics. Anticoagulation has been used in some patients, but its utility is currently unclear.

A 67 year-old man with past medical history of hypertension and aortic stenosis was referred from an outside hospital due to hematemesis. Isolated gastric varices in the fundus demonstrating red wale sign were found on upper endoscopy at the outside hospital. The patient underwent triple phase computed tomography (CT) of the abdomen and pelvis, which showed moderate intrahepatic bile duct dilation in segment 4 and compression on the intrahepatic left portal vein concerning for neoplasm. Other notable findings included a non-cirrhotic-appearing liver and a massively dilated appendix. MRI confirmed presence of an ill-defined region of enhancement in segment 4 as well as a 3.2 x 2.3 cm mass in the pancreatic tail. Endoscopic retrograde cholangiopancreatography was performed with plastic stent placement into the area of intrahepatic bile duct obstruction. Endoscopic ultrasound-guided fine needle aspiration of the pancreas mass was performed, with cytology positive for adenocarcinoma. The patient later underwent repeat upper endoscopy with successful gastric variceal embolization using a mixture of histoacryl and lipiodol. He was scheduled for exploratory laparotomy with biopsy of peritoneal nodules, partial left hepatectomy and appendectomy. Pathology was positive for adenocarcinoma in all specimens, with special stains indicating pancreatic origin. The patient was discharged home with plan to initiate palliative chemotherapy. Pancreatic adenocarcinoma is associated with poor long-term survival, in part due to non-specific symptoms leading to delay in diagnosis until the disease is advanced. Metastasis is common, most frequently to the liver, peritoneum and lungs. Metastasis from the pancreas to the appendix is very rare, with only a couple of case reports in the literature. There are case reports documenting an association between pancreatic cancer and gastric varices, but typically the splenic vein is thrombosed, causing left sided portal hypertension. In this unusual case, the splenic vein was patent on contrast-enhanced CT and MRI. Since gastric variceal bleeding is life-threatening, prompt diagnosis and treatment are essential. Control of bleeding is most commonly performed using medication to reduce portal pressure and either glue injection into varices or transjugular intrahepatic portosystemic shunt placement. Malignancy should be considered in the differential diagnosis for patients with gastric variceal bleeding without known history of cirrhosis.Figure: Gastric varices.Figure: Gastric varices after histoacryl/lipiodol injection.

Introduction: Percutaneous endoscopic gastrostomy (PEG) is a relatively safe procedure for long-term enteral feeding, with procedure-related mortality rates as low as 0 to 2%. One of the complications after PEG tube placement is inadvertent tube removal, which occurs in 1.6 to 4.4% of cases. A Foley catheter (FC) is often placed temporarily by the emergency department or nursing facility to maintain gastrostomy tract patency. We report a case of acute pancreatitis in a patient with a FC balloon that migrated into the jejunum. Case presentation: A 59-year-old woman presented from a nursing home with acute-onset abdominal pain and vomiting. Her medical history was significant for traumatic brain injury, with resulting dysphagia leading to PEG tube placement. Several weeks prior to presentation, the gastrostomy tube was replaced by a FC due to inadvertent dislodgement. Physical exam at presentation revealed epigastric abdominal tenderness. The gastrostomy stoma showed no sign of erythema or infection, but the length of FC outside the patient's body was only approximately 10 cm. Laboratory studies showed elevated serum lipase of 5923 Units/L and a clinical diagnosis of pancreatitis was made. An upper abdominal ultrasound showed a normal gallbladder without cholelithiasis or bile duct dilation. Serum triglyceride level was normal. Abdominal CT with IV contrast revealed acute uncomplicated pancreatitis. The FC tip and a dilated balloon were seen in the proximal jejunum. Since no alternative cause of pancreatitis was found, it was postulated that due to FC's lack of an internal retention mechanism, distal migration of the balloon caused obstruction at the ampulla of Vater, resulting in acute pancreatitis. The balloon was deflated and the FC was repositioned so that the internal balloon was relocated to the stomach. When the patient clinically improved, an upper endoscopy was performed and the FC was replaced by a low profile gastrostomy button. Discussion: Acute pancreatitis is a rare complication after a FC is used to replace a standard gastrostomy tube. A FC should be substituted only on a temporary basis and active medical intervention should be sought to place a new gastrostomy tube. While the FC is in place, care should be taken to periodically inspect the length of tubing external to the body. In patients with temporary gastrostomy tubes, migrated catheters should be considered in the differential of patients presenting with acute pancreatitis.

Introduction: Portal vein thrombosis (PVT) typically occurs in patients with cirrhosis and other prothrombotic disorders. Among patients with cirrhosis, PVT is common and correlates with severity of the patient's liver disease. The prevalence of PVT is estimated to be less than 1 percent in patients with compensated cirrhosis, but increases to 8 to 25% in liver transplant candidates. Hepatocellular carcinoma (HCC) frequently occurs in the setting of cirrhosis, and transcatheter arterial chemoembolization (TACE) is one of many non-surgical treatment options available to treat HCC. There is extensive literature on safety and efficacy of TACE for HCC in the setting of pre-existing PVT. However, development of PVT after TACE has not been previously described in the literature. This study aims to review the incidence of PVT in patients after transcatheter arterial embolization procedures at a single tertiary care, liver transplantation center. Methods: The study included all patients who underwent TACE or bland particle embolization from 1/1/2013 to 12/31/2015 at our institution. The study population included a total of 217 cases that underwent embolization procedures for HCC. Pre-treatment, procedural and post-treatment imaging was reviewed to evaluate characteristics of the initial lesion (i.e. size, location), assess for pre-existing PVT and diagnose new PVT. Results: Of 217 cases reviewed, a total of 19 (8.8%) procedures had new diagnosis of PVT on follow-up imaging. Sixteen patients were found to have new PVT and 3 had probable PVT with no subsequent confi rmatory imaging. Twelve patients had hepatitis C (HCV) cirrhosis (63%), three had alcoholic cirrhosis (16 %), two had mixed alcoholic and HCV cirrhosis (10.5%) and two were diagnosed with cryptogenic or NASH cirrhosis. Drug-eluting beads loaded with doxorubicin were used in 18 of the PVT-associated procedures, while one patient had bland embolization resulting in PVT. Fourteen cases were diagnosed using magnetic resonance imaging and four cases were diagnosed using computed tomography. Conclusion: Our findings support the possibility that TACE may be an inciting event for developing new PVT. Further research needs to be performed to determine whether PVT is a complication of TACE or a sequela of advancing cirrhosis or cancer itself. If further studies do demonstrate that PVT can be a complication of TACE, this finding would likely have a significant impact on the management of HCC patients as they await liver transplantation.