Upper GI Bleeding: Nonvariceal UGIB

0:00Welcome to Board Pearls. This is episode one of two of the Upper GI Bleeding chapter, in the Stomach and Small Bowel Disorders module. This episode is non-variceal upper GI bleeding, and the thing to hold onto is that it's a sequence, and the order is fixed for a reason: it's a perfusion problem first, a diagnosis problem second, and an endoscopy problem third. Stabilize the patient, optimize the field, risk-stratify, then scope. The first hour moves mortality more than the procedure does, so anything that jumps the order surrenders the leverage the early steps actually have.

0:36The opening moves are two large-bore IVs with a type and crossmatch going, and a monitored bed when the risk score predicts severe disease. Intubation is liberal in ongoing hematemesis, altered mental status, or advanced encephalopathy, because aspiration is the leading preventable cause of death in massive bleeding. And labs including lactate matter, because a high lactate or base deficit flags the patient who's sicker than the blood pressure is admitting.

1:04The transfusion threshold is the first place the mechanism teaches the rule: transfuse red cells at a hemoglobin of seven in most patients, and the reason a restrictive target beats a liberal one is instructive. Pouring in blood expands the plasma volume, and in a cirrhotic that raises portal pressure and pops varices that were holding, so the liberal strategy does worse, most clearly in the cirrhotic. The exceptions come from the patient, not the bleed: transfuse sooner for active hemodynamic instability because the lab number lags the patient, and transfuse at a higher target around eight to ten in acute coronary syndrome because the ischemic myocardium doesn't tolerate anemia and the cardiac event would be the thing that kills them, with chronic stable coronary disease a softer reason to aim for eight. The trap is the stable cirrhotic with a variceal bleed where you're tempted to transfuse up to nine or ten; that's exactly the patient the data show gets worse with the higher target.

1:56Acid suppression before endoscopy rests on one piece of pharmacology that runs through the whole chapter: the intragastric pH targets are layered. Above four prevents stress-related damage, above five slows the clot from being digested, and above six inactivates the enzyme that lyses clot and lets platelets build a durable plug. Below six the clot keeps dissolving as it forms. A high-dose intravenous proton pump inhibitor, an eighty-milligram bolus then an infusion, drives the pH above six and holds it there. Pre-endoscopy, that downstages the lesion, so fewer high-risk findings are seen and fewer patients need endoscopic therapy at that procedure, but once endoscopic therapy has been delivered, the pre-endoscopy dose doesn't further change rebleeding or mortality. The guideline couldn't firmly recommend for or against it, but the favored answer when it's offered against no acid suppression is to start it, because it downstages lesions and the scope is sometimes delayed. And erythromycin given half an hour or so before the scope earns its place by clearing the field: it's a motilin agonist that sweeps gastric blood and clot into the duodenum so the endoscopist finds the lesion on the first pass, reducing repeat procedures even though it doesn't change transfusion or length of stay.

3:14Risk stratification separates who needs admission and urgent endoscopy from who can go home for an outpatient scope. The Glasgow-Blatchford score uses only pre-endoscopy data, and its value is at the low end: a score of zero or one identifies a very-low-risk patient who almost never needs intervention, so the guideline endorses outpatient management with prompt outpatient endoscopy for them. That's where the real question lives, whether anyone needs admission at all. AIMS65 predicts ICU need and mortality but isn't the tool for finding the go-home patient, and the Rockall score needs the endoscopy itself to compute, so it's the wrong pick when the stem asks who needs to be admitted, because the data aren't in hand yet.

3:55Timing of endoscopy is the next decision, and the answer is non-variceal endoscopy within twenty-four hours, with no advantage to going before six. Taking the sickest patients to the suite urgently didn't improve rebleeding or mortality and actually led to more intervention, because the clots hadn't stabilized and the lesions looked worse than they would have a few hours later. Variceal bleeding is the exception, with a tighter target, and that's the next episode. And one thing not to do: tranexamic acid is not used in acute upper GI bleeding; the large trial settled it, so don't pick it as an adjunct even when the stem hints at coagulopathy.

4:32The second part is the anticoagulated and antiplatelet patient, organized by one principle: reversal is agent-specific and resumption is time-specific. Reverse only what needs reversing, dose by mechanism, and resume as soon as bleeding control allows, because what kills these patients over the next two weeks is the clot the anticoagulant was preventing, not the rebleed. Warfarin in significant bleeding is reversed with four-factor prothrombin complex concentrate rather than plasma, and the reason is mechanical: the concentrate corrects the INR within minutes, in a small volume that won't flood an elderly cardiac patient, delivering the clotting factors directly, whereas plasma works too slowly and adds a real volume load. You add intravenous vitamin K to hold the reversal past the half-life of those factors, and picking plasma first-line is wrong unless the concentrate is unavailable. Dabigatran is reversed with idarucizumab, an antibody fragment that binds the drug far more tightly than the drug binds thrombin, neutralizing it within minutes, with dialysis as a backup since dabigatran is dialyzable. The factor Xa inhibitors, apixaban and rivaroxaban, are reversed with andexanet alfa, a decoy that mops up circulating drug, reserved for life-threatening bleeding because of cost and a thrombotic concern, with off-label prothrombin complex concentrate as an alternative when it's unavailable.

5:58Antiplatelet management starts from why the patient is on the drug, because holding aspirin isn't free. Aspirin for secondary cardiovascular prevention is continued through the bleed unless it's life-threatening, and if it was held, it's resumed the day hemostasis is confirmed; the trap is the clean-base ulcer in a patient whose aspirin was held for three weeks during the workup, because that patient has been unprotected too long. Aspirin for primary prevention is a different calculation and is generally stopped for good after a bleed. The dual-antiplatelet patient is managed with cardiology, because a recent drug-eluting stent will thrombose if the therapy is interrupted in the first month, so the move is to drop to a single agent during active bleeding when the indication allows and resume the second as soon as hemostasis is confirmed. And platelet transfusion is the wrong answer when the platelet count itself is fine, because transfusing platelets to reverse an antiplatelet drug showed a harm signal and no benefit; you give platelets for an actual low count or before a procedure, not as routine reversal. On resumption, warfarin is typically restarted within a week of hemostasis, sometimes within a few days in atrial fibrillation once the lesion is treated, and the direct anticoagulants similarly, because the late mortality is thrombotic.

7:11The third part is the endoscopy itself, and the Forrest classification maps what the endoscopist sees to what they do. The single idea is that high-risk stigmata get dual therapy because their rebleed rate is high, and low-risk stigmata get nothing because it's too low to justify. The categories: a spurting arterial bleed and an oozing bleed without a discrete vessel are both active hemorrhage and both get treated; the non-bleeding visible vessel, a pigmented protuberance that's the eroded artery, has a high enough rebleed rate that it also gets treated; the adherent clot that resists vigorous irrigation is the intermediate, debated one; and the flat pigmented spot and the clean ulcer base have low rebleed rates, so they're observed, with the clean-base patient often discharged. So only the active bleeders and the visible vessel clearly mandate intervention.

8:03Dual therapy is the favored approach for those high-risk lesions, and epinephrine alone is explicitly not adequate, because its tamponade and vasoconstriction wear off and the vessel rebleeds. So epinephrine injection is the first half, buying a bloodless field, and the second half is what actually achieves durable hemostasis: a thermal method like bipolar or heater probe, or a clip. The choice depends on the lesion, with clips better in fibrotic or thin-walled bases where heat risks perforation and thermal better in soft, accessible ulcers where the probe can be pressed on cleanly. The adherent clot is the genuinely controversial slot, where the guideline is intentionally permissive: you can either snare the clot off and treat whatever high-risk stigma is underneath, which the data modestly favor, or manage with high-dose acid suppression alone, and the caveat is the deep penetrating or hard-to-reach ulcer, where snaring can turn a stable picture into a bleed you can't control.

8:59A couple of other tools. Hemostatic powder spray, which adheres to a wet bleeding surface, is a temporizing or salvage measure that washes off within a day or two, so it bridges to definitive therapy rather than standing alone, favored for the actively bleeding ulcer that's failed conventional therapy, a field obscured by heavy bleeding, or a diffusely oozing tumor. And the over-the-scope clip is the rescue for recurrent ulcer bleeding after an initial successful hemostasis, outperforming another round of standard therapy, so the favored answer in the recurrent bleeder, once you've characterized the lesion, is the over-the-scope clip rather than a third standard attempt or a jump to surgery.

9:36Certain lesions predict that hemostasis will fail, and they're worth knowing because they change the upfront plan: the posterior duodenal bulb ulcer erodes into the gastroduodenal artery and the lesser-curve gastric ulcer into the left gastric artery, both high-rebleed locations, and along with ulcers over two centimeters, instability at presentation, and active bleeding at the scope, they flag the patient who needs aggressive dual therapy and an early parallel conversation with interventional radiology and surgery. When endoscopic therapy fails or the patient rebleeds, the next step is repeat endoscopy, not immediate surgery, because a second attempt often succeeds and spares an operation; embolization comes after a failed second endoscopy, and surgery sits behind embolization. Routine second-look endoscopy after a successful hemostasis isn't recommended, reserved for suspected rebleeding or an inadequate first exam.

10:31The fourth part is the medical treatment that turns a good procedure into durable control, and the dosing follows that pH pearl from earlier. After successful hemostasis of a high-risk lesion, you give seventy-two hours of high-dose intravenous acid suppression, either a continuous infusion or intermittent high dosing, which are equivalent, keeping the pH above six so the clot matures and the rebleed risk drops, then transition to oral therapy twice daily and taper over several weeks to let the ulcer heal, taken before meals because the drug only works on an active pump. Low-risk lesions, the flat spot and clean base, don't need the infusion at all; they go straight to oral therapy and most are discharged quickly. Vonoprazan, the newer acid blocker that gives a higher, flatter suppression without needing an empty stomach, is the favored alternative when the standard drug has failed or a faster, more sustained effect is needed, though high-dose intravenous acid suppression remains the default.

11:29H. pylori eradication is mandatory in every bleeding peptic ulcer and is one of the most underweighted secondary-prevention steps, because eradication produces a larger drop in the yearly rebleed rate than any other single intervention. You test at the index endoscopy, but with a specific trap: blood in the stomach lowers the sensitivity of biopsy-based tests, so a positive is trustworthy but a negative may be false, which is why serology is useful here given the high pre-test probability, and a breath or stool test off acid suppression can be used later, with eradication confirmed weeks after treatment off the acid blocker. And the discharge plan finishes with NSAID counseling: stop the NSAID if it caused the ulcer, and if anti-inflammatory therapy is essential, switch to a COX-2 selective drug at the lowest dose with an acid blocker, resume secondary-prevention aspirin within a week with acid-blocker cover, and keep gastroprotection going as long as antiplatelet therapy continues. One follow-up trap: a gastric ulcer gets a repeat scope at six to twelve weeks, especially if large or inadequately biopsied, because a cancer can masquerade as a benign-looking ulcer that heals symptomatically and comes back as cancer, whereas duodenal ulcers don't need that routine follow-up.

12:45So the whole episode is a sequence tested in order. The first hour is perfusion, with a hemoglobin target of seven and a cardiac exception near eight. The pre-endoscopy phase adds acid suppression to downstage the lesion and erythromycin to clear the field, and the Glasgow-Blatchford score identifies who can go home, with endoscopy within twenty-four hours and no gain from going earlier. The anticoagulated patient gets agent-specific reversal, prothrombin complex concentrate for warfarin, idarucizumab for dabigatran, andexanet reserved for life-threatening Xa-inhibitor bleeding, and early resumption because the late risk is clotting. The endoscopy runs on the Forrest classification, dual therapy for high-risk stigmata and nothing for low-risk, with the over-the-scope clip as the rescue for the recurrent bleeder. And the medical follow-through is seventy-two hours of high-dose acid suppression after a high-risk hemostasis, then oral therapy, plus mandatory H. pylori testing and eradication in every ulcer bleeder.

13:44Episode two takes the other side of the chapter: acute variceal hemorrhage on its bundle of a vasoactive drug, antibiotics, and band ligation, with early TIPS for the high-risk patient, and then the unusual causes, from Dieulafoy and Mallory-Weiss through GAVE, Cameron lesions, hemobilia, and the herald bleed of an aortoenteric fistula.

14:05For 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 two of chapter nine, and I'll see you in the next one.