Upper gastrointestinal bleed

This edition of Morning Report focuses on a fictional patient with an upper gastrointestinal bleed.

The patient

Chief symptom: Dark stools for the past three days.

Photo by Thinkstock
Photo by Thinkstock.

History: A 67-year-old man with a history of type 2 diabetes mellitus and osteoarthritis presented to the ED with reports of dark stools for three days associated with lightheadedness. He reported inability to maintain his usual schedule due to fatigue. He also reported his stools to appear darker than usual with a sticky texture and a malodorous smell. He described a recent worsening in intensity of a long-standing epigastric burning sensation, which he had experienced intermittently for years. He reported doubling his usual dose of calcium carbonate without any significant relief. He had no associated chest pain, palpitations, dyspnea, or syncope. His current medications included metformin, 500 mg twice a day; aspirin, 81 mg/d; and ibuprofen, 400 mg as needed for pain. He reported no surgical or pertinent family history. He reported a 20 pack-year smoking history but no alcohol or illicit drug abuse.

Physical examination: Vitals were as follows: supine blood pressure, 120/80 mm Hg; supine heart rate, 110 beats/min; standing blood pressure, 90/60 mm Hg; standing heart rate, 120 beats/min; temperature, 98.6 °F; respiratory rate, 20 breaths/min; and oxygen saturation, 100% on ambient air. The patient appeared anxious and diaphoretic with moist skin. Conjunctival pallor was noted on ocular exam. Cardiovascular and respiratory examinations were within normal limits. Abdominal examination revealed normoactive bowel sounds with mild epigastric tenderness to palpation and no evidence of hepatosplenomegaly or masses. Rectal exam revealed normal perineal exam, normal sphincter tone, no external hemorrhoids or rectal masses, and a positive result on a fecal occult blood test.

Labs: A complete blood count (CBC) revealed a white blood cell count of 9,000 cells/mm3, platelet count of 210,000 U/L, hemoglobin level of 7.9 g/dL, hematocrit of 23%, and mean corpuscular volume of 90 fL. Prothrombin time, international normalized ratio, and partial thromboplastin time were all normal. Blood urea nitrogen (BUN) level was 45 mg/dL, and creatinine level was 1.0 mg/dL. Liver function tests were normal. Amylase and lipase levels were normal. Fecal occult blood testing was positive.

Imaging: Electrocardiogram (EKG) revealed sinus tachycardia at 110 beats/min with no overt ischemic changes. Chest and abdominal X-rays were within normal limits. CT of the abdomen without oral contrast revealed mild ileitis.

Clinical course

The patient was admitted to the step-down unit. He was started on a continuous IV of pantoprazole infusion with IV maintenance fluids and continued NPO. Gastroenterology was consulted for further evaluation. A repeat hemoglobin level, rechecked the following morning, was 7.2 g/dL. He subsequently underwent an upper endoscopy, which demonstrated an actively bleeding visible vessel in the gastric wall. Epinephrine injection was endoscopically administered, which resolved the bleeding. No other bleeding was noted. After the procedure, the patient started a clear liquid diet. IV medications were discontinued, and oral pantoprazole, 40 mg/d, was started. The following morning, he reported feeling better. He was discharged home on oral pantoprazole and told to follow-up with primary care in two weeks.

Approximately one week posthospitalization, the patient returned to the ED, obtunded with severe hematemesis and respiratory distress. His family reported continued dark stools since discharge, which had worsened to multiple episodes of hematochezia on the day of readmission. Vital signs were as follows: heart rate, 140 beats/min; blood pressure, 80/40 mm Hg; temperature, 99.2 °F; respiratory rate, 11 breaths/min; and oxygen saturation, 79% on ambient air. Examination revealed an obtunded patient who was not following commands, with evidence of bright red blood in the oropharynx and severe epigastric tenderness to palpation. Urgent labs revealed a hemoglobin of 4.8 g/dL and a BUN of 51 mg/dL. The patient was emergently intubated and started on mechanical ventilation. He was started on IV normal saline boluses, IV pantoprazole, and two units of packed red blood cells. During blood transfusion, the patient had a cardiac arrest with no return of spontaneous circulation after 30 minutes of resuscitation, and he was pronounced dead.

Case review

Q: Was our patient diagnosed correctly?

A: Yes, the patient was diagnosed correctly for a suspected upper gastrointestinal (GI) bleed, and gastroenterology was consulted for endoscopy to evaluate the source of the bleeding.

Q: Was our patient managed and discharged appropriately?

A: Our patient was managed appropriately initially. He was admitted and immediate resuscitative efforts with IV fluid hydration via large-bore IVs began while he was awaiting endoscopy. He was started on IV pantoprazole therapy with monitoring of hemoglobin and hematocrit. However, blood transfusion should have been initiated in the scenario of active bleeding and hypovolemia.

Endoscopy was performed within 24 hours to discover the bleeding source. Assessment and therapy of the bleeding site were done on time. However, epinephrine injection therapy should be used in combination with clipping or coaptive coagulation in cases of high-risk stigmata bleeding (1). Aspirin should have been withheld and resumed only after revaluation of the benefits and risks of such therapies. Helicobacter pylori testing should have been performed in the case of peptic ulcer bleeding.

Another error was the premature discharge of the patient. Patients who receive endoscopic treatment for high-risk stigmata should be hospitalized for 72 hours to monitor for rebleeding, since most rebleeding occurs during this time (1). Proton-pump inhibitors (PPIs) should have been administered intravenously for 72 hours or, if the patient remained stable, twice-daily IV PPI could have been considered to practice cost-conscious care (2).

Q: Were care transitions handled appropriately at discharge?

A: No, our patient should have been advised to follow up one week after discharge to check hemoglobin and hematocrit and to seek immediate medical attention if his symptoms worsened. Discharge counseling should have been provided regarding risk of rebleeding, as well as avoidance of NSAIDs and aspirin. Our patient should have been advised that melena can continue for several days postprocedure but should decrease. He should have been told to watch for symptoms such as weakness, palpitations, and dyspnea and to return for immediate care if symptoms worsened.


Upper GI bleeding results in high patient morbidity and mortality and has an annual incidence of around 100 per 100,000 adults (1). Anatomically, it is defined as bleeding from a source proximal to the ligament of Treitz. It is broadly categorized into variceal and nonvariceal bleeding. Nonvariceal bleeding is more common than variceal, with peptic ulcer being the most common cause of hospitalization (3). Other causes of upper GI bleeding are erosive lesions of the esophagus, stomach, or duodenum, as well as vascular and traumatic causes. The hospitalization rate for upper GI bleeding is estimated to be sixfold higher than for lower GI bleeding (4). Common risk factors associated with bleeding peptic ulcers are H. pylori infection, NSAID or aspirin use, physiologic stress, and excess gastric acid.

Common clinical features of upper GI bleeds include hematemesis and melena. Hematochezia can also occur in the setting of a massive bleed. Patients with severe blood loss and hypovolemia may present with syncopal episodes. Taking a good medical and medication history is paramount in patients with suspected upper GI bleeding. Any prior episode of bleeding, history of liver disease complicated by varices, peptic ulcer disease with a history of H. pylori infection, or NSAID use may suggest potential bleeding sources and precipitating factors. NSAIDs have been implicated as an important factor in nonhealing ulcers (5).

On physical examination, clinicians should look for signs of hypovolemia like resting tachycardia or orthostatic hypotension and signs of active bleeding like presence of frank bloody emesis. Rebound tenderness should be assessed in patients with acute abdominal pain to rule out perforation. The initial work up should include a complete blood count, comprehensive metabolic panel, coagulation profile, blood typing, and cross matching for packed red blood cells, followed by checking of hemoglobin and hematocrit every four to six hours. Initial hemoglobin levels are a poor indicator of the true extent of severity of bleeding. As patients experience blood loss, the effect of this blood loss won't be reflected in the hemoglobin level immediately; it will be hours before a change in hemoglobin levels is seen. Therefore, it is best to rely on signs of hemodynamic instability to direct therapy.


Assessment of airway, breathing, and circulation is the topmost priority. Protection of the airway with intubation may be needed to avoid respiratory compromise from potential aspiration of blood and gastric contents, especially in patients with active bleeding and altered mental status (6). All patients who present with signs and symptoms of upper GI bleeding should be evaluated immediately for hemodynamic stability and managed accordingly by rapid intravascular volume replacement with isotonic crystalloid fluids (7). It has been demonstrated that early and aggressive resuscitation reduces mortality (8). After initial hemodynamic resuscitation, risk stratification based on clinical, laboratory, and endoscopic features is recommended by the International Consensus Upper Gastrointestinal Bleeding Conference Group (1). Prognostic scales such as Blatchford and Rockall scores are recommended for early stratification of patients into low- and high-risk categories for rebleeding and mortality.

The Glasgow Blatchford Score (GBS) (Table 1) is recommended for pre-endoscopy risk stratification and is based on patients' admission hemoglobin level, BUN level, heart rate, and systolic blood pressure, as well as presentation with syncope or melena, and evidence of hepatic disease or cardiac failure (9). The score ranges from zero to 23, and the risk of requiring blood transfusion and endoscopic intervention increases with a higher score. It is suggested that patients younger than age 70 years with a score of 2 or lower be managed as outpatients. A GBS above 7 at admission is associated with higher rebleeding rates following discharge (10). The Rockall score, which is a better predictor of mortality, uses endoscopic variables along with age, shock, and major comorbidities (11).

All patients with hemodynamic instability or active bleeding should be admitted to the intensive care unit for resuscitation and close observation. All patients should be kept NPO to facilitate urgent endoscopy. Nasogastric lavage is useful for removing particulate matter, fresh blood, and clots from the stomach to facilitate endoscopy. A gastroenterologist should be consulted for all suspected GI bleeding.

A PPI is empirically administered intravenously to all patients before endoscopy to downstage the bleeding lesion. The goal is to keep the intragastric pH above 6 to stabilize the clot. It may also reduce the need for endoscopic therapy (12). PPI therapy with or without endoscopic hemostasis for bleeding ulcers showed a significant and consistent reduction in the risk of recurrent bleeding and the need for surgery (13). The European Society of Gastrointestinal Endoscopy recommends initiating high-dose PPI via IV bolus followed by continuous infusion (80 mg, then 8 mg/h). However, PPI infusion should not delay the performance of an early endoscopy. Somatostatin is used in the treatment of variceal bleeding and is not routinely recommended for patients with acute ulcer bleeding, though it may reduce the risk of continued bleeding due to nonvariceal causes (14).

If there are clinical signs of significant blood loss, blood transfusion should be considered immediately. International guidelines recommend starting blood transfusion when hemoglobin levels are below 7.0 g/dL and targeting hemoglobin levels of 7.0 g/dL to 9.0 g/dL in the absence of acute hemorrhage, tissue hypoperfusion, and myocardial ischemia (15, 16). In patients with active bleeding, hemodynamic instability, and other comorbidities such as coronary artery disease, the threshold hemoglobin level for transfusion may be higher (1). A single dose of IV erythromycin (250 mg) 30 to 120 minutes before endoscopy significantly improves the endoscopic visualization and reduces the need for repeat endoscopy (17).

When possible, anticoagulants and antiplatelet agents should be held in patients with upper GI bleeding. The thrombotic risk of reversing anticoagulation should be weighed against the risk of continued bleeding for each patient, based on individual cardiac and gastrointestinal risks (18).

Endoscopic treatment

Endoscopy has become the mainstay for diagnosis and treatment of acute upper GI bleeding and is recommended within 24 hours of presentation. Endoscopic therapy should be undertaken for ulcers with high-risk stigmata to reduce the risk of rebleeding. Endoscopy within 12 hours may be considered in patients with hemodynamic instability that persists despite attempted volume resuscitation, in-hospital bloody emesis/nasogastric aspirate, or contraindication to the stoppage of anticoagulation as it may improve outcomes (7).

The Forrest classification is widely used to differentiate between low- and high-risk endoscopic stigmata in patients with peptic ulcer bleeding (19). Type Ia and Ib and type IIa and II b are classified as high-risk stigmata and warrant endoscopic hemostatic therapy. For type IIb with an adherent clot, endoscopic clot removal by vigorous irrigation is helpful to expose the ulcer bed, and the lesion should be managed either by hemostatic or intensive PPI therapy according to the findings. Endoscopic hemostatic therapy is not routinely indicated for patients with low-risk stigmata, including type IIc and type III.

Commonly used endoscopic hemostatic modalities are injection therapy (epinephrine, sclerosing agents, and adhesive agents), thermal therapy (contact and non-contact), mechanical therapy (clips and band ligation), and topical therapy. Epinephrine injection therapy should be used in combination with other hemostatic modalities, and this combination method is recommended for high-risk stigmata peptic ulcer bleeding (20-22).

Postendoscopic management

High-dose antisecretory therapy with an IV bolus followed by continuous infusion of a PPI (80 mg, then 8 mg/h) is the recommended pharmacological treatment. It should be continued for 72 hours postendoscopy for patients with high-risk stigmata. Low-dose therapy, such as a twice-daily IV bolus of PPI, can also be used for 72 hours postendoscopy. High-dose PPI therapy in a high-risk group significantly reduces the risk for rebleeding, surgery, and mortality (23).

Repeat or second-look endoscopy can be considered in selected patients with high risk of bleeding or with clinical evidence of rebleeding. Surgery or transcatheter angiographic embolization is only recommended when a second attempt at endoscopic hemostasis fails (24). H. pylori should be tested for and treated in all patients with bleeding secondary to a peptic ulcer. Patients with an initial negative test in the acute setting should be retested, and documentation of successful eradication of H. pylori is strongly recommended (1).

Patients who need long-term anticoagulation should be restarted on anticoagulant therapy in a timely manner. Recommended timing is based on patient risk factors, and resuming warfarin between 7 to 15 days is comparatively safe and effective (25). Aspirin for primary cardiac prophylaxis should be withheld, and cardiology consultation is recommended to evaluate the risks and benefits of restarting aspirin. Discontinuation of low-dose aspirin can significantly increase risk of death and acute cardiovascular events (26). Aspirin for secondary cardiac prophylaxis should be withheld for three days after endoscopy in those with high risk of rebleeding and immediately resumed along with PPI in those with low risk of rebleeding (7, 27). The need for NSAIDs should be carefully evaluated in patients with NSAID-associated bleeding ulcers. For those who need NSAIDs, a cyclooxygenase-2 selective NSAID like celecoxib is recommended at the lowest effective dose, along with a PPI (28).

Discharge management

All patients should be discharged on an oral PPI for a specific duration, and an appointment with primary care physician in the first week after discharge is recommended. In patients with high-risk stigmata endoscopic findings, two weeks of PPI twice daily followed by once daily for at least six to eight weeks with follow-up with primary care and gastroenterology is recommended. Patients with low-risk stigmata endoscopic findings should receive PPI once daily. In patients with H. pylori infection, triple or quadruple therapy treatment should be followed by PPI alone for another two weeks (2).

Multidisciplinary teams should be employed to smooth the transition of care. Some patient factors that predict readmission due to rebleeding are shock at presentation; melena; age older than 60 years; associated comorbidities like heart, liver, or renal failure; larger ulcer size; and stigmata of recent hemorrhage on endoscopy (11). Some factors in care that predict readmission include early discharge within 72 hours for high-risk patients, inadequate PPI dosage, and insufficient discharge advice.


Appropriate documentation is essential to facilitate continuity of care, communication among clinicians, and appropriate billing and reimbursements. For upper GI bleeding due to a gastric ulcer, it is important to specify the primary diagnosis as acute and/or chronic gastric ulcer with hemorrhage. Complications and comorbidities should also be documented along with the cause to allow coding to accurately reflect severity and to ensure appropriate diagnosis-related group classification (Table 2) (29).

GI bleeding is self-limiting in over 80% of cases. Accurate risk stratification can identify low-risk patients who might not require hospitalization or who could be discharged soon after admission, increasing the cost-effectiveness of their care.

In summary, this case demonstrates the potential opportunities for better management of patients with upper GI bleeding, including adhering to guideline-directed therapy, improving the transition of care, and providing proper documentation, which can result in better patient outcomes and fewer readmissions due to rebleeding.