GI Emergencies: GI Emergencies: Massive Upper GI Bleed

0:00Welcome to Board Pearls. This is episode one of four of the GI Emergencies chapter, in the Special Populations and Acute or Supportive Care module. This episode is a massive upper gastrointestinal bleed, from the door to definitive control. The first patient is a man on warfarin vomiting blood, and his case shows why a massive bleed is a resuscitation problem before it's an endoscopy problem, how much blood to give and when to switch to a massive transfusion protocol, how to reverse warfarin and whether reversal has to finish before the scope, which drug goes in before endoscopy and which never does, when the airway is secured, how soon to scope, and what happens when the endoscopist can't stop the bleeding. The second patient is a woman with an old aortic graft and a warning bleed, and her case shows the cause that can't be missed.

0:45Here's the patient. A seventy-one-year-old man takes warfarin for atrial fibrillation and ibuprofen for back pain. He arrives vomiting large volumes of bright red blood. His heart rate is one hundred twenty-eight and his blood pressure eighty-two over fifty. His hemoglobin is six point nine and his international normalized ratio, INR, three point four. He's alert but frightened, and he vomits again in the bay. He has no liver disease and no history of aortic surgery.

1:15He has a massive upper gastrointestinal bleed: hematemesis with shock, on an anticoagulant, and his ibuprofen makes a peptic ulcer the likely source. A massive upper bleed is a resuscitation problem before it's an endoscopic one, because the decisions that drive outcome happen in the first hour. Those decisions are airway protection, restrictive transfusion, a massive transfusion protocol if bleeding continues, drug-specific reversal of his anticoagulant, and preparing his stomach for endoscopy. The full management of ulcer and variceal bleeding is in chapter nine, and this episode covers the time-critical decisions that set a massive bleed apart.

1:56His first decision is how much blood to give, and the answer is less than instinct suggests. In a large randomized trial of acute gastrointestinal bleeding, a restrictive strategy, transfusing below a hemoglobin of seven grams per deciliter with a target of seven to nine, lowered six-week mortality and rebleeding compared with a liberal strategy of transfusing below nine. Patients with cirrhosis drove most of the benefit, because liberal transfusion expanded portal volume and raised variceal pressure. The reason restriction works is that overzealous resuscitation reopens partly clotted vessels by raising mean arterial pressure, and it dilutes the platelets and clotting factors the patient is laying down at the bleeding site.

2:38Two qualifications change that threshold. Active acute coronary syndrome raises it to eight, because the same anemia that protects a clot starves an ischemic heart, so if he had chest pain with ST-segment depression, his target would be eight. And a patient visibly exsanguinating is transfused on clinical grounds rather than on a laboratory value, because the measured hemoglobin lags real loss by about thirty to sixty minutes, so the number drawn is the number from before the latest surge. His six point nine is already below seven, but with ongoing hematemesis and a pressure of eighty-two, he's transfused now on what's visible, without waiting for the next result.

3:17If he keeps bleeding despite that, he needs a massive transfusion protocol. It's triggered when bleeding continues despite initial resuscitation, with an anticipated need of more than ten units of red cells in twenty-four hours or four units in one hour. The protocol delivers red cells, plasma, and platelets in a balanced one-to-one-to-one ratio, which reproduces whole blood and avoids the dilutional coagulopathy that crystalloid or red cells alone produce. The ratio comes from the trauma literature, and it applies here because the physiology is the same. Calcium is checked and replaced, because citrate in stored products binds ionized calcium, which the clotting cascade needs, and ionized calcium falls with rapid transfusion. After three units in forty minutes he's still vomiting blood with a pressure that won't hold, on track for four in an hour, so the protocol is activated and his calcium replaced.

4:12His warfarin has to be reversed at the same time, not afterward. Patients on warfarin or a direct oral anticoagulant, DOAC, who bleed massively need drug-specific reversal alongside transfusion. Warfarin in life-threatening bleeding is reversed with four-factor prothrombin complex concentrate, 4F-PCC, rather than plasma. Vitamin K is added for sustained reversal, but it isn't the answer for the acute window, because its onset takes twelve to twenty-four hours. Dabigatran is reversed with idarucizumab, five grams intravenously. Apixaban, rivaroxaban, and edoxaban are reversed with 4F-PCC, because andexanet alfa was withdrawn from the United States market in December twenty twenty-five.

4:59His reversal doesn't have to finish before his endoscopy. Reversal is paired with definitive intervention rather than being a reason to delay it. Endoscopic hemostasis can be performed with an INR of one point five to two point five while reversal proceeds, and an INR above two point five is reversed before endoscopy in most situations. His INR of three point four means 4F-PCC and intravenous vitamin K now, with a recheck before the scope. An INR of two point one wouldn't have held up his endoscopy.

5:32Two drugs are considered before his endoscopy, and only one is given. He's already receiving an intravenous proton pump inhibitor, an eighty-milligram bolus of pantoprazole or esomeprazole followed by eight milligrams an hour, as chapter nine covers. Erythromycin, two hundred fifty milligrams intravenously over twenty to thirty minutes, given thirty to ninety minutes before endoscopy, is a motilin agonist that empties blood and clot from the stomach so the endoscopist can see the source. It reduces the need for a second-look endoscopy and shortens the hospital stay, without a mortality benefit. Tranexamic acid isn't given in acute upper bleeding. A large randomized trial showed no benefit and more venous thromboembolism, because upper bleeding is dominated by arterial spurting from a discrete vessel, which needs mechanical hemostasis, not the fibrinolysis-driven coagulopathy that antifibrinolytics are designed to treat. He gets erythromycin, not tranexamic acid.

6:23His airway is secured before the scope, because endoscopy will bring up more blood. Intubation is indicated for grade three to four hepatic encephalopathy, massive hematemesis with aspiration risk, hemodynamic collapse, or any patient who can't protect the airway during the high-volume blood return expected at endoscopy. An encephalopathic patient with cirrhosis and active hematemesis, for instance, is intubated electively before esophagogastroduodenoscopy, EGD, rather than as an emergency with the patient positioned and blood in the field. He's alert, but he's vomiting large volumes, which is massive hematemesis with aspiration risk, so he's intubated first.

7:01How soon he's scoped depends on the kind of bleed and whether he can be stabilized. For non-variceal bleeding, endoscopy is within twenty-four hours. In a randomized trial of high-risk patients with a Glasgow-Blatchford score of twelve or more, endoscopy within six hours didn't improve mortality, rebleeding, or the need for intervention compared with endoscopy at six to twenty-four hours. Variceal bleeding gets endoscopy within twelve hours, as part of the bundle of octreotide, ceftriaxone, band ligation, and early transjugular intrahepatic portosystemic shunt, TIPS, for Child-Pugh C disease or Child-Pugh B with active bleeding at the first endoscopy. Endoscopy under six hours is reserved for ongoing massive hemorrhage that can't be stabilized, a suspected aortoenteric fistula, or other urgent concerns. He won't stabilize, so he's scoped as soon as he's intubated, reversed, and resuscitated enough to tolerate it.

7:58His endoscopy finds the source and treats it. It shows a large posterior duodenal bulb ulcer with an actively spurting vessel. High-risk stigmata like his get dual therapy: epinephrine injection plus a second modality, thermal coagulation, a hemoclip, or absolute ethanol injection. Hemostatic powder is used for diffuse oozing or malignancy, and an over-the-scope clip for refractory rebleeding. His vessel is injected and clipped, and the bleeding stops. Because his ulcer was the cause of his bleed, his ibuprofen is stopped if at all possible.

8:32His ulcer's location is also its weakness, because the posterior bulb is one of the two sites where endoscopic therapy most often fails. Salvage angiography by interventional radiology is the next step when dual endoscopic therapy fails. The arteries most successfully embolized are the gastroduodenal artery, for posterior duodenal bulb ulcers, and the left gastric artery, for ulcers on the lesser curve of the stomach. Those are the same two sites where endoscopy most often fails, because both are fed by branches running just beneath a thin ulcer floor, and posterior bulb ulcers rebleed in roughly half of cases. Six hours after his endoscopy he rebleeds, and he goes to angiography for gastroduodenal artery embolization. Surgery is reserved for bleeding that continues despite endoscopy and embolization, a suspected aortoenteric fistula, and free perforation.

9:26One cause of upper bleeding can't be missed, because its first bleed stops. A seventy-six-year-old woman had an abdominal aortic aneurysm repaired with a graft six years ago. Three days ago she had a single self-limited episode of hematemesis, and today she has melena. She has a suspected aortoenteric fistula. It classically presents with a herald bleed, a self-limited bleed days to weeks before a catastrophic hemorrhage, in a patient with a prior aortic graft. A fistula can be primary, from a mycotic aneurysm, or secondary, from a graft, and it communicates with the third part of the duodenum in up to seventy-five percent of cases. Computed tomography, CT, angiography, looking for periaortic gas or contrast extravasation, comes before both EGD and surgical repair, and this is one of the situations where endoscopy within six hours is justified. Her bleed stopping is a warning, not a reassurance, and a confirmed fistula goes to surgery.

10:28Here are the traps the boards set in these patients. Transfusion targets a hemoglobin of seven, not nine or ten. An exsanguinating patient is transfused without waiting for a laboratory value. Active coronary ischemia raises the threshold to eight. Massive hemorrhage needs plasma and platelets with the red cells, plus calcium. Warfarin is reversed with 4F-PCC rather than plasma, vitamin K alone is too slow, and andexanet is no longer available. Endoscopy isn't delayed to correct an INR between one point five and two point five. Tranexamic acid isn't given. Massive hematemesis means securing the airway before the scope. A stable patient doesn't need endoscopy within six hours. And a herald bleed in a patient with an aortic graft is never reassuring.

11:17So, start to finish. Resuscitate before the scope. Transfuse to a hemoglobin of seven, or eight with active coronary ischemia, and on clinical grounds if the patient is exsanguinating. Activate massive transfusion with one-to-one-to-one products and calcium when bleeding is torrential. Reverse anticoagulants by drug and in parallel: 4F-PCC and vitamin K for warfarin, idarucizumab for dabigatran, and 4F-PCC for the factor ten A inhibitors, reversing an INR above two point five before endoscopy.

11:52Give intravenous erythromycin before endoscopy, never tranexamic acid, and intubate first if the airway is threatened. Scope non-variceal bleeding within twenty-four hours and variceal bleeding within twelve, treating high-risk stigmata with dual therapy. When endoscopy fails, embolize, and operate when embolization fails too. After an aortic graft, a herald bleed means CT angiography first.

12:20Episode two is acute lower GI bleeding and acute mesenteric ischemia.

12:25For the full chapter, the practice vignettes, and the topic-tagged question bank, head to board pearls dot com. You'll find the rest of the series on Apple Podcasts, Spotify, or wherever you listen to podcasts. That brings us to the end of episode one of four of chapter thirty-one, and I'll see you in the next one.