Next, we will have Dr. Anna DeLoy. Dr. DeLoy is an assistant professor and therapeutic endoscopist at the University of Colorado, and she will be discussing the management of portal hypertensive bleeding. Good afternoon. My name is Anna DeLoy, and I'm going to be talking about esophageal and gastroecvariceal bleeding. I have no disclosures. Case one is a 63-year-old with a history of compensated cirrhosis who presents with melanoma. It's important to stress that any upper GI bleed in a patient with cirrhosis or portal hypertension should be managed as a variceal bleed until proven otherwise. Esophageal varices are the most common type of GI varices. Their prevalence ranges from 42 to 75% in patients with child's class A through C. They bleed in 5 to 15% of patients per year, and the six-week mortality after index bleeding is as high as 25%. Endoscopy should be performed in any case of suspected variceal bleeding within 12 hours of admission. Delaying endoscopy is a risk factor for increased mortality and should be avoided. This is a list of things that should be done prior to your EGD. Esophageal hemorrhage is a medical emergency, and patients should be admitted or transferred to the ICU. If possible, I prefer that my patients are intubated prior to EGD. Patients should be resuscitated with a target hemoglobin of 7 to 9. This has shown a lower mortality than a liberal transfusion strategy. Antibiotics should be given to any cirrhotic who is bleeding. This decreases the risk of bacterial infections and increases survival. Ceftriaxone 1 gram per day is most commonly used and typically continued for 7 days. Basoactive agents like octreotide decrease acute bleeding, transfusion requirements, and mortality. Octreotide is initially given as a 50-microgram bolus, followed by a 50-microgram-per-hour infusion. This should be started prior to endoscopy and continued for about 3 to 5 days after endoscopy. Bacterial ligation, or banding, is first line for the treatment for bleeding esophageal varices. It has better hemostasis, a lower rate of side effects, and a lower rate of early mortality when compared to sclerotherapy. Banding should be started at the GE junction. However, of course, if there is an actively bleeding varix or even a white nipple sign, which is a platelet-fibrin plug at a site of recent bleeding, then a bandage should first be placed on those sites. Use bursus suction to make sure the varix will suction into the cap. The red-out sign, as shown in the bottom-left picture, occurs when the varix is fully sucked into the cap and the band can then be fired. It's good to remember that the second-to-last band will be a different color. All right, back to our case. Our patient is scoped and successfully banded. Four hours later, you get a call that the patient is having hematemesis. You repeat an EGD and find ongoing bleeding that cannot be stopped despite placing additional bands. You call IR for a TIPS, but they are unavailable for several hours. What do we do now? Balloon tamponade and stents can both be used as a bridge to TIPS. This picture depicts the three tubes most commonly used. Blakemore and Minnesota tubes have an esophageal and a gastric balloon, whereas a Linton tube, seen on the left side of the picture, only has a large gastric balloon. The rate of hemostasis with balloon tamponade is high, however, re-bleeding occurs in more than half of patients after deflation. Complications from balloons include esophageal ulceration or perforation and aspiration pneumonia. Balloons should not be left inflated for greater than 24 hours, and most who use Blakemore or Minnesota tubes don't inflate the esophageal balloon to minimize the risk of complications. Stents are fully covered and can be placed endoscopically without the need for fluoro. They can achieve hemostasis in 80 to 96 percent of cases and can be left in place for up to two weeks. Their major limitation is that they require advanced training in order to place and therefore limits their applicability for all endoscopists. A small multicenter randomized controlled trial compared metal stents with balloon tamponade in patients with refractory esophageal variceal bleeding and showed a higher hemostasis rate and fewer adverse events with stents. However, the six-week survival rate was not significantly different, and again, a limitation of stents is that not all endoscopists are trained in their placement. To summarize, banding is first line for bleeding esophageal varices. If this fails or there is recurrent bleeding that can't be stopped endoscopically, a TIPS is indicated. Esophageal stents or balloon tamponade can be used as a temporary bridge to TIPS. All right, let's move on to case two. This is a 50-year-old with a history of alcoholic cirrhosis, hepatic encephalopathy, chronic panc, and splenic vein thrombosis who presents after hematemesis at home. Gastric varices are less common than esophageal varices but are associated with higher morbidity and mortality. The serine classification is used for risk and treatment stratification of gastric varices. GOV type 1 are the most common gastric varices and are esophageal varices extending into the lesser curvature. GOV type 2 are esophageal varices extending into the fundus. IGV type 1 are isolated varices in the fundus, and IGV type 2 are isolated varices located elsewhere in the stomach. GOV type 2 and IGV type 1 are considered gastric fundal varices. Hemodynamically, esophageal varices and GOV1s arise from the left and right gastric veins, whereas IGV1s and GOV2s are usually supplied by the short and posterior gastric veins. Gastric fundal varices are often accompanied by a gastrorenal or a splenorenal shunt. Factors associated with a high risk of bleeding are location, size, and presence of high-risk stigmata. Management of gastric varices is not well-defined the way it is for esophageal varices. Treatment options for bleeding gastric varices primarily include endoscopic and IR-guided therapies. GOV1 varices are treated similarly to esophageal varices and can be banded. Endoscopic options for gastric fundal varices include direct endoscopic injection of glue and EUS-guided treatments. The benefit of direct injection is that it doesn't require any specialized treatment or advanced training. However, it is unlikely to be a good option in very large fundal varices, and it's difficult to perform in cases where there's brisk active bleeding leading to poor visualization. The benefit of EUS-guided treatment is that it allows you to target and detect the vessel lumen regardless of whether there is active bleeding. You can control delivery of treatment and confirm treatment success. However, a major limitation is that these techniques require training in EUS and therefore may not be practical in many centers. Based on the current data, EUS-guided coil plus glue has shown lower bleeding rates, transfusion requirements, re-intervention rates, and mortality when compared with direct endoscopic glue injection. Dual therapy with EUS-guided coil plus glue has been shown to be safer with lower re-bleeding and re-intervention rates than with monotherapy. IR-guided options for bleeding gastric varices include TIPS, balloon-occluded retrograde transvenous obliteration, and coil-assisted retrograde transvenous obliteration. For both BRTO and CRTO, a portosystemic shunt such as a gastro-renal or splenorenal shunt must be present. A catheter is advanced up the jugular or femoral vein as depicted in this little picture into the left renal vein and then into the varix outflow tract or the shunt. The varix is occluded with a balloon-tipped catheter followed by delivery of coils and or sclerosis. This procedure can increase portal pressures and therefore it can worsen complications such as ascites or bleeding from esophageal varices. It's important to note that based on current AASLD guidelines, TIPS is the treatment of choice for IgV1 and GoV2 varices. The decision on which treatment should be used depends on institutional and endoscopic expertise, variceal location, i.e., GoV1 can be banded but fundal varices cannot, the presence of portal hypertension, shunts, and thromboses, obtaining cross-sectional imaging with CTA is important for determining the vascular anatomy prior to determining which treatment might be a good option, and comorbidities also affect treatment. Since TIPS can worsen encephalopathy, BRTO can worsen esophageal varices, and also ascites, one might not want to choose a TIPS for someone who has severe hepatic encephalopathy. Since the management of GVs is still evolving and not well-defined in the guidelines, here are two different examples of treatment algorithms at different institutions. It's important to stress once again that one of the biggest factors in determining treatment is currently institutional expertise. Management of GoV1 varices is more straightforward and these should be banded by GI. The algorithm on the left from Dr. Botts' group is primarily based on EOS-guided therapies with IR as a salvage therapy. The treatment algorithm on the right uses IR's first line if a shunt is present and endoscopic therapies if a shunt is absent. For centers without GI or IR expertise, temporary rescue measures that can be used for bleeding GVs include balloon tamponade with a linten tube and or hemo spray. Here's a third algorithm for bleeding gastric varices, and we're going to go through this, but it's important to emphasize that direct comparisons between different endoscopic and EOS methods are sparse and limited in sample size. Meta-analyses comparing different methods suggest EOS-guided therapy is superior to direct endoscopic therapy. Additional studies looking at different EOS modalities suggest that EOS-guided coil plus glue injection has the best efficacy in the treatment of gastric varices. However, it's very important to note that we need additional large, well-designed RCTs to confirm all of these potential benefits. So, going through this algorithm here, we're going to focus on the right side of the screen where you look at acute GV bleeding. EOV1, like we discussed before, treat them as if they're esophageal varices with endoscopic band ligation. EOV2 or IGV1, the next fork in the algorithm is whether or not there's EOS expertise. So, if there's not EOS expertise, which you see on the left side, then options include endoscopic glue injection or IR modalities like Birdo or TIPS. If you can't achieve hemostasis with that, you see at the bottom, just as we mentioned before, potential sort of backups are Lenten tube, because it has a bigger gastric balloon than a Blakemore or a Minnesota tube. Potentially splenectomy, that's if a splenic vein thrombosis caused the gastric varices or potentially hemo spray. But again, we need additional studies to confirm this. On the right side, if you do have EOS expertise at your institution, then options include obviously EOS modalities such as coil injection plus glue. If that gives you hemostasis, great. If not, then you go back onto the left side and options that include endoscopic glue injection, BRTO or TIPS. I'm going to go through the endoscopic treatment modalities now. In direct glue injection, the needle is primed with water. The needle is then inserted directly into the varix and one milliliter of glue is rapidly injected, followed by about one milliliter of water to clear the remaining glue from the dead space of the catheter lumen. The needle should then be promptly extracted to prevent it from being embedded into the varix and water should be passed into the catheter lumen at pretty high speeds for about 15 to 20 seconds to prevent closure of the catheter lumen. Injection can be repeated until the varix is completely obliterated and there is a feeling of hardness when you probe it. Some endoscopic start injections at the side of the varix as opposed to the variceal dome due to lower pressures and less chance of bleeding. The most commonly injected glues are 2-octyl cyanoacrylate or N-butyl cyanoacrylate. 2-octyl has a longer polymerization time. Lipidol slows polymerization, which decreases the chances of the glue hardening within the catheter but can also increase the chances of glue embolization. Adverse events in glue injection include systemic embolization, needle impaction within the varix, and ulcer formation. We're going to walk through the steps of EOS-guided treatment. Antibiotics such as Cipro should be given during the procedure and for five days post-procedure. The patient should be placed supine with a left lateral tilt to help preferentially fill the gastric fundus with water. Inject water, usually a couple of hundred cc's to fill the fundus. Position the linear EOS scope in the distal esophagus or fundus. Carefully examine the varices to determine which vessels are intramural and which are extramural. Sometimes the feeder vessel and shunts can also be identified. Extramural vessels and shunts should be avoided as targets. Here in this video, we can see the arrow pointing out the intramural vessels and the lumen of the fundus below it. The arrow is now on what's thought to be a feeder vessel and this can be traced by rotating the scope clockwise and counterclockwise extramural. You can see the spleen there for a second and now back towards intramural. Both 19 gauge and 22 gauge FNA needles can be used. 19 gauge needles allow for bigger coils. The needle should be primed with saline and then advanced into the varix. Aspirated blood to confirm your needle is within the varix. This video created by Dr. Ben Muller shows how coils are loaded. The varix has been punctured and blood is aspirated. The needle is then flushed with saline. The coil introducer is inserted and lure locked to the FNA needle hub. Advance the length of coil into the FNA needle using the stylet as a pusher. And then remove the introducer and stylet. And finally reintroduce the stylet to advance the coil through the needle into the varix. The use of fluoroscopy after coil placement is variable. Some endoscopists inject contrast to confirm the absence of runoff and exclude a shunt prior to injecting glue to further reduce the risk of embolization. Common brands of coils are Nestor and MRI. If the GI department doesn't have them, they can typically be borrowed from IR. They come in a range of diameters and lengths. A 22-gauge needle can only accommodate diameters less than 10 millimeters. So if you are treating a large varix, you will need a 19-gauge needle so that a larger diameter coil can be used. The length of the coil determines how many loops it will form. And typically the number of coils will depend on the size of the varix being treated. After coil placement, glue can then be injected. Most people are using 2-octyl cyanoacrylate or Dermabond without lipidol. Other groups are injecting a gelatin sponge slurry to further minimize the risk of embolization. The glue is followed by 1 milliliter of normal saline to flush it completely out of the needle. Doppler is then used to confirm a reduction in flow. Notably, number of coils, which coils, the use of fluoro and type of glue is variable between providers and we don't yet know which is superior. Our patient gets a CTA and is found to have a large splenic vein thrombosis. There is no evidence of a shunt. EGD shows a large bundled varix with signs of recent bleeding. After a discussion with IR, TIPS is felt to be a poor choice because of her history of hepatic encephalopathy and a BRTO is not an option given the lack of a shunt. Because of the large size of the varix and the location of the varix, the decision is made to proceed with EOS guided coiling plus glue injection. Doppler confirms flow within the varix, the variceal nest is measured, 19-gauge needle punctures into the varix, and a coil is injected. The second coil is then deployed given the large size of the varix. Glue is then injected and Doppler is used to confirm a reduction in flow consistent with successful treatment. In summary, banding is superior to sclerotherapy for the management of esophageal varices. TIPS should be used for refractory bleeding in esophageal varices. Treatment for GVs is rapidly evolving. You should consider patient comorbidities, vascular anatomy, the varix location, and institutional endoscopist expertise when deciding the best treatment. I'd like to acknowledge and thank these doctors for their help with and contributions to the material in this presentation. Thank you.

portal hypertensive bleeding

esophageal variceal bleeding

cirrhosis

endoscopy

banded ligation