Perdarahan Gastrointestinal

01Pathophysiology

Gastrointestinal (GI) bleeding is divided anatomically by the ligament of Treitz (the suspensory ligament of the duodenum at the duodenojejunal junction). Bleeding proximal to this landmark is classified as upper GI bleeding (UGIB), and bleeding distal to it is classified as lower GI bleeding (LGIB). This distinction is not arbitrary; it dictates the diagnostic approach, the initial test of choice, and the management pathway.

In peptic ulcer disease, the most common cause of UGIB, disruption of the gastric or duodenal mucosa by acid and pepsin erodes into the submucosal vasculature. Posterior duodenal ulcers are particularly dangerous because they can erode into the gastroduodenal artery, producing brisk arterial hemorrhage. Gastric ulcers along the lesser curvature can erode into the left gastric artery. The mechanism is straightforward: mucosal defense (bicarbonate secretion, prostaglandin-mediated blood flow, mucus barrier) is overwhelmed by aggressive factors (acid, NSAIDs, H. pylori). NSAIDs inhibit cyclooxygenase (COX), reducing prostaglandin synthesis and thereby weakening the mucosal barrier. H. pylori directly damages epithelial cells and provokes chronic inflammation.

In variceal bleeding, the pathophysiology begins with portal hypertension, most commonly from liver cirrhosis. As hepatic architecture is distorted by fibrosis, resistance to portal blood flow increases. Portal pressures above 12 mmHg (measured by the hepatic venous pressure gradient, or HVPG) place the patient at risk for variceal hemorrhage. Blood is shunted through portosystemic collaterals, including the left gastric (coronary) vein, which feeds esophageal submucosal venous plexuses. These thin-walled, dilated veins rupture under high pressure, producing sudden, voluminous hematemesis. This is why variceal bleeding tends to be dramatic and immediately life-threatening.

A Mallory-Weiss tear results from a longitudinal mucosal laceration at the gastroesophageal (GE) junction, typically caused by forceful or prolonged retching and vomiting. The mechanism is sudden increases in intra-abdominal pressure transmitted to the GE junction. The tear involves the mucosa and submucosa but does not penetrate the full thickness of the wall (which would instead be a Boerhaave syndrome, a transmural esophageal rupture, a critical distinction on exams).

For lower GI bleeding, diverticulosis is the most common cause of significant hematochezia in older adults. Diverticula form at points where the vasa recta (nutrient arteries) penetrate the colonic wall. Over time, the artery at the dome of the diverticulum becomes exposed and can rupture, producing painless, large-volume bright red or maroon rectal bleeding. This is arterial bleeding, which is why it can be brisk despite originating from a small structural lesion.

Angiodysplasia (arteriovenous malformations) is the second most common cause of LGIB in the elderly. These are degenerative, dilated, thin-walled vascular structures in the submucosa, most often found in the right colon (cecum and ascending colon). The mechanism involves chronic, low-grade venous obstruction of submucosal veins as they pass through the muscular layer, leading to progressive dilation and eventual formation of arteriovenous communications. There is a known association between angiodysplasia and aortic stenosis (Heyde syndrome), where acquired von Willebrand factor deficiency from shear stress across a stenotic valve contributes to bleeding.

The reason melena (black, tarry, foul-smelling stool) is characteristic of UGIB is that blood is exposed to gastric acid and intestinal bacteria during transit, converting hemoglobin to hematin and other oxidized pigments. As little as 50 to 100 mL of blood in the upper tract can produce melena. Hematochezia (bright red blood per rectum) usually indicates a lower GI source, but in the setting of hemodynamic instability, it can also indicate a brisk upper GI source where transit is too rapid for oxidation to occur. This is a commonly tested concept.

02Classification and Clinical Manifestation

The Forrest classification is clinically important because it determines whether endoscopic therapy is indicated. Forrest Ia, Ib, and IIa lesions require endoscopic intervention (such as thermal coagulation, hemoclipping, or injection therapy). Forrest IIb (adherent clot) is controversial, but current guidelines favor removing the clot and treating the underlying lesion. Forrest IIc and III do not require endoscopic therapy and can be managed medically with PPI alone.

03Diagnostic Workup

The Best Initial Test for any patient presenting with GI bleeding is clinical assessment and resuscitation, combined with basic laboratory studies (CBC, BUN/Cr, coagulation panel, type and crossmatch). These do not tell you the source, but they establish the severity and guide transfusion decisions.

An important nuance: hemoglobin and hematocrit may be normal early in acute bleeding because blood loss is whole blood (both red cells and plasma are lost proportionally). It takes 24 to 72 hours for equilibration with extravascular fluid to reveal the true drop. Serial measurements are therefore essential. This is a commonly tested concept: do not be falsely reassured by a normal initial hemoglobin in an acutely bleeding patient.

A BUN/Cr ratio greater than 20:1 is a useful clue pointing toward an upper GI source. Blood in the upper tract is digested like a protein meal; amino acids are absorbed in the small intestine and converted to urea in the liver, raising the BUN disproportionately to creatinine. This is a soft sign, not a definitive test, but it frequently appears in vignettes as a discriminating detail.

Nasogastric aspirate can be helpful when the source is uncertain. A bloody or coffee-ground aspirate confirms an upper source. However, a clear, non-bilious aspirate does not rule out UGIB because a duodenal source may not reflux into the stomach, especially if the pylorus is closed. A clear, bilious aspirate makes an upper source much less likely.

For confirmed or suspected UGIB, upper endoscopy (EGD) is both the best initial diagnostic test and the gold standard. It should be performed within 24 hours of presentation. In patients with suspected variceal bleeding or hemodynamic instability despite resuscitation, endoscopy should be performed within 12 hours or even more urgently. EGD allows direct visualization, biopsy, and therapeutic intervention in the same session.

For LGIB, colonoscopy is the best initial diagnostic and therapeutic test. The patient typically requires bowel preparation first, and colonoscopy is ideally performed within 24 hours. In patients with hemodynamically significant LGIB where colonoscopy is not feasible (too brisk, patient too unstable), CT angiography is the next step. CTA can localize the bleeding site and guide subsequent angiographic embolization.

Tagged RBC scan is the most sensitive noninvasive test for detecting GI bleeding. It can identify bleeding at very low rates (0.1 mL/min). However, it has poor anatomic localization, so it is most useful for confirming that active bleeding is occurring and roughly localizing it before angiography.

Mesenteric angiography requires a higher bleeding rate to be diagnostic ($\ge \backslash geq$ 0.5 mL/min), but it offers the advantage of therapeutic embolization. It is typically reserved for patients who are bleeding too briskly for colonoscopy or who have failed endoscopic therapy.

04Management and Treatment

Acute stabilization follows the ABCs. Two large-bore peripheral IVs (18-gauge or larger) should be placed immediately. Volume resuscitation with isotonic crystalloid is initiated while blood products are being prepared. The current evidence supports a restrictive transfusion strategy with a target hemoglobin of $\ge \backslash geq$ 7 g/dL in hemodynamically stable patients. Over-transfusion in variceal bleeding is particularly harmful because it can raise portal pressures and worsen hemorrhage. In patients with active coronary disease, the transfusion threshold is generally higher (8 to 9 g/dL).

For nonvariceal UGIB, intravenous PPI therapy is the cornerstone. Pantoprazole 80 mg IV bolus followed by 8 mg/hr continuous infusion for 72 hours is the standard regimen. The rationale is that clot stability over an ulcer bed requires a gastric pH above 6; continuous high-dose PPI achieves this. After 72 hours, the patient is transitioned to oral PPI (pantoprazole 40 mg twice daily, then once daily). All patients with peptic ulcer disease should be tested for H. pylori and treated if positive (triple therapy: PPI + clarithromycin 500 mg twice daily + amoxicillin 1 g twice daily for 14 days, or a bismuth quadruple regimen). NSAIDs must be discontinued.

For variceal UGIB, three interventions are started simultaneously before endoscopy: (1) octreotide (a somatostatin analog that reduces splanchnic blood flow and portal pressure), given as a 50 mcg IV bolus followed by 50 mcg/hr infusion for 3 to 5 days; (2) prophylactic antibiotics (ceftriaxone 1 g IV daily for 7 days), which have been shown to reduce bacterial translocation, spontaneous bacterial peritonitis, rebleeding, and mortality; (3) urgent endoscopy with endoscopic variceal ligation (banding), which is preferred over sclerotherapy. If banding fails to control hemorrhage, a TIPS procedure is the next step. Balloon tamponade (Sengstaken-Blakemore or Minnesota tube) is a temporizing bridge measure, not definitive therapy, and carries risks of esophageal necrosis and aspiration.

Primary prophylaxis of variceal bleeding is indicated in cirrhotic patients found to have medium or large varices on screening endoscopy. Options include nonselective beta-blockers (propranolol or nadolol, titrated to a resting heart rate of 55 to 60 bpm) or endoscopic variceal ligation. Cardioselective beta-blockers (such as metoprolol) are not effective for this purpose because they do not block the beta-2 receptors responsible for splanchnic vasoconstriction. Secondary prophylaxis (after a first variceal bleed) uses the combination of a nonselective beta-blocker plus EVL.

For lower GI bleeding, most diverticular hemorrhages resolve spontaneously without intervention. For persistent or recurrent bleeding, colonoscopy with endoscopic therapy (hemoclipping, thermal coagulation, epinephrine injection) is the first-line approach. If colonoscopy fails or bleeding is too brisk, angiographic embolization via interventional radiology is pursued. Surgery (segmental colectomy) is reserved for patients who fail all other measures.

05Differential Diagnosis and Distractors

06Traps and High-Yield Pearls

The most common way students get questions on GI bleeding wrong is by failing to consider an upper GI source in a patient presenting with hematochezia. Exam vignettes will describe a hemodynamically unstable patient with bright red blood per rectum, and the student reflexively selects colonoscopy. The correct answer is to evaluate the upper tract first (EGD or at minimum NG aspirate) because a brisk upper source can easily present with hematochezia. The rule to remember: hemodynamic instability plus hematochezia should trigger evaluation for UGIB before assuming a lower source.

Another frequently tested trap is the initial hemoglobin. Students see a normal hemoglobin in an acutely bleeding patient and conclude the bleeding is not significant. Early hemoglobin is unreliable because acute whole blood loss does not immediately change the concentration; hemodilution has not yet occurred. Serial hemoglobin measurements and clinical signs (tachycardia, hypotension, orthostasis) are far more reliable indicators of severity.

For variceal bleeding, a critical pearl is that antibiotics are not optional. Students often select octreotide and endoscopy but forget that prophylactic antibiotics (ceftriaxone) are a standard part of the initial management bundle and independently reduce mortality. Another trap: the question may ask about the first step in a known cirrhotic with hematemesis. The answer is resuscitation and stabilization, not endoscopy. Endoscopy is urgent but comes after the patient is hemodynamically supported.

In the differential between Mallory-Weiss and Boerhaave, the discriminator is transmural perforation. The vignette for Boerhaave includes chest pain, subcutaneous emphysema, or pneumomediastinum on imaging. Mallory-Weiss involves hematemesis after vomiting but no perforation signs. Students confuse these two regularly.

For lower GI bleeding in children, always think of Meckel diverticulum when the presentation is painless rectal bleeding in a child under 2. The classic teaching mnemonic is the "rule of 2s": 2% of the population, 2 feet from the ileocecal valve, 2 inches long, presents by age 2, two types of ectopic tissue (gastric and pancreatic). The diagnostic test is a Technetium-99m pertechnetate scan (Meckel scan), which detects ectopic gastric mucosa.

Finally, the aortoenteric fistula is the diagnosis students miss with the worst consequences. Any patient with a prior abdominal aortic aneurysm repair who presents with GI bleeding, even if the initial bleed is small, must be evaluated for this entity. The "herald bleed" concept is a classic exam scenario: a small, self-limited GI bleed in a patient with an aortic graft is not benign and should prompt emergent CT angiography and surgical consultation.

The core competency being tested across GI bleeding questions is the ability to localize the source (upper vs lower), choose the correct initial diagnostic and therapeutic modality, and recognize life-threatening causes that require emergent intervention rather than routine workup.