Management of upper gastrointestinal bleeding

Acute UGIB is classified into two broad categories viz non-variceal and variceal, with the former being more common and including oesophagitis, oesophageal ulcers, oesophageal cancer, MalloryWeis tear, gastritis, gastric ulcer, gastric cancer, Dieulafoy’s lesion, vascular ectasias (angiodysplasia), duodenal ulcer, aorto-enteric fistula, haemobilia and haemosuccus pancreaticus.8 South African Family Practice 2019; 61(5):15-18

The pathophysiology of non-variceal bleeding involves either arterial haemorrhage as in ulcers and mucosal deep tears, or low-pressure venous haemorrhage as seen in telangiectasis and angioectasis. 2 Variceal causes, accounting for 10% of cases, include bleeding from gastric varices, oesophageal varices or portal hypertensive gastropathy. 9,10 The underlying mechanism is thought to be raised portal pressure transmitted to oesophageal and gastric varices leading to portal gastropathy. 2 Peptic ulcer disease is still the most common cause of UGIB and is significantly related to Helicobacter pylori infection. 2 This organism disrupts the mucous barrier and has a direct inflammatory effect on gastric and duodenal mucosa. H. pylori colonises > 50% of the population, with 10-20% of colonised individuals becoming symptomatic and developing ulcers.
The goal of assessment (through appropriate history, physical examination, laboratory tests), is to determine the severity of the bleed, recognise potential sources of the bleed and search for conditions that may influence definitive management. 11

Clinical presentation
Presenting complaints may include weakness, dizziness, nausea or dyspepsia. The presence of haematemesis localises bleeding to the foregut. Frank, red blood is an indication of moderate to severe ongoing bleeding while coffee-ground vomitus suggests a delayed bleed. Coffee-ground vomitus is formed by blood that has oxidised within the stomach and is darkened and thickened.
A large haemorrhage in the region of the foregut may manifest as haematochezia (red or maroon stool) with rapid transit of blood through the gastrointestinal tract usually associated with orthostatic hypotension. Melaena on the other hand, is black tarry stool suggestive of bleeding that originates proximal to the Ligament of Treitz. Those presenting with frank haematemesis tend to have more severe bleeding than those with just melaena. 11 Recurrence of bleeding from the same lesion is common and a past medical history must be determined. In addition, a history of dyspepsia with nocturnal symptoms, early satiety, peptic ulcer disease, malignancy, angiodysplasia, nonsteroidal antiinflammatory drug use, and previous aortic surgery must be elicited.
Preexisting co-morbid conditions like coronary artery disease, pulmonary disease, renal failure, coagulopathies, thrombocytopaenia, hepatic dysfunction and dementia must be ascertained as these conditions may affect patient management.

Physical examination
Physical examination will determine the haemodynamic status of the patient. Significant examination findings of a tachycardia above 100 beats per minute, systolic blood pressure less than 90 mmHg, increased capillary refill time above two seconds and cold peripheries are early warning signs of decreased perfusion.
Determination of orthostatic hypotension indicates a blood loss of at least 20% of the blood volume. 11 The patient must be examined for stigmata of chronic liver disease like spider angiomata, gynaecomastia, collateral abdominal veins, ascites and asterixis. The tell-tale abdominal and groin scars indicative of previous aortic surgery might point Table I. Important points to consider in the history and examination of a patient with UGIB Important points in history taking to assess the site, severity and aetiology of UGIB

Examination must include Investigations
General symptoms -fatigue, weakness, dizziness, syncope The page number in the footer is not for bibliographic referencing www.tandfonline.com/oemd 17 to the presence of a herald bleed (aorta-enteric fistula) and urgent need for computed tomographic angiography (CTA).
Although nasogastric tube placement is controversial, a positive aspirate in the presence of haematochezia is suggestive of an upper gastrointestinal tract bleed. A negative aspirate on the other hand does not rule out upper gastrointestinal bleeding. Aspiration of red blood or coffee ground is predictive of an acute lesion or an inactive bleed respectively (Table I).

Laboratory investigations
The basic investigations that will determine and guide further management include a full blood count, urea and electrolytes, liver function tests and coagulation studies.
Normocytic red blood cells will suggest acute bleeding while a microcytic appearance or iron deficiency is indicative of chronic blood loss. A high urea to creatinine ratio is expected in cases of UGIB due to absorption as blood passes through the gastrointestinal tract, possibly exacerbated by decreased renal perfusion. 12

Risk stratification
Risk scores utilising clinical factors and laboratory data are routinely used to predict adverse outcomes in UGIB. 13 The Rockall score and the Glasgow-Blatchford score are the two widely used stratification tools that predict the likelihood of continuation of bleed or of re-bleeding and the risk of death. 14,15 The Glasgow-Blatchford score (0-23 points) does not consider endoscopic criteria and is useful in the emergency room. 11 This scoring system helps in selecting which patients may benefit from outpatient endoscopy (Table II).
The Glasgow-Blatchford score seems to be superior to the preendoscopic Rockall score in all important clinical outcomes including mortality. The recently developed CANUKA score was found to be superior at identifying low-risk patients and performed similar in predicting re-bleeding, therapeutic intervention and mortality. 14

Emergency management
The principles of management for all patients with UGIB are similar. Standard care should commence with ensuring that the airway is open and patent, breathing is adequate and that the circulatory system is adequate for perfusion. If there is an altered level of consciousness, the airway needs to be secured and protected. All patients should receive supplemental oxygen via a nasal canula. 6 Intravenous peripheral access with two wide bore catheters must be obtained early, followed by immediate fluid resuscitation with crystalloids, either Ringers lactate or normal saline. Fluid balance must be strictly monitored following insertion of a Foleys catheter.
In the haemodynamically unstable patient, resuscitation should continue with packed red blood cells (RBCs). In patients with suspected variceal bleeding, caution must be taken against over transfusion as this can precipitate worsening of the bleeding. 11 The goal of transfusion should be aimed at maintaining a haemoglobin level greater than 7 g/dL except for patients with unstable coronary artery disease who are at risk for adverse events. In the setting of anaemia -the goal should be maintaining a haemoglobin level above 9 g/dL. 11 Acid suppression with intravenous proton pump inhibitors is initiated and reviewed immediately after the source of bleeding is identified. The European Society of Gastroenterology (ESGE) recommends using high-dose PPIs given as a bolus (e.g. omeprazole 80 mg IVI) and then running an infusion over 72 hours (e.g. omeprazole 8 mg/hr). 16 However, systemic review comparing intermittent versus continuous infusion therapy proved similar efficacy, thus therapy can be tailored for each patient. 17,18 Intravenous erythromycin (3 mg/kg over 30 min) if given one to two hours prior to endoscopy will improve visualisation and diagnostic and therapeutic yield at the time of endoscopy. Metoclopramide is an alternative prokinetic. 11 Early upper endoscopy within 24 hours of patient presentation is the diagnostic modality of choice for UGIB. It is used to establish the bleeding site and also achieve haemostasis. Endoscopic modalities for achieving haemostasis include self-expandable metallic stents and endoscopic variceal ligation for oesophageal varices; cyanoacrylate injection for gastric varices; and endoscopic haemospray for nonvariceal UGIB. Interventional radiological techniques such as balloon-occluded retrograde transvenous obliteration and transjugular intrahepatic portosystemic shunting can be considered for gastric varices.

Conclusion
As family practitioners, we may be the first medical contact for a patient who presents with UGIB. Appropriate resuscitation based on risk stratification must be aimed at restoring haemodynamic stability and stopping the bleeding. Early endoscopic therapy within 24 hours is required for most patients. The Glasgow-Blatchford score identifies a subset of patients suitable for outpatient management. Consideration on subsequent management following an UGIB should focus on prevention of re-bleeding.