The presentation and severity of GI bleeding is as diverse as the aetiology. This article discusses emergency presentation requiring immediate hospital transfer. More chronic conditions presenting with similar symptoms warrant urgent referral to the outpatient gastroenterology or colorectal clinic.
Assessment and management
Assessing GI bleeding in the community can be difficult. The focus of assessment should be on identifying patients with haemodynamic compromise and those with risk factors for poor outcome.
A simultaneous 'ABC' approach is required in the acutely unwell patient, with heart rate and blood pressure used for simple assessment of hypovolaemia.
More subtle signs include tachypnoea, oliguria, anxiety, and confusion. Advancing age, and comorbidities such as cardiac failure, malignancy, and liver disease are risk factors for mortality.
The presence of haematemesis, haematochezia, or gross rectal bleeding (such as clots per rectum) is also associated with poor outcome. Immediate transfer to hospital should be arranged and IV access and fluid resuscitation should be initiated.
|Common causes of gastrointestinal bleeding|
|Upper GI||Lower GI|
Upper GI bleeding
The incidence of acute upper GI (UGI) bleeding in the UK is 90-170 per 100,000 adults per year, and accounts for 50,000 admissions to hospital.1,2 The overall mortality is 7-10%. UGI bleeds are classified as non-variceal (the majority) or variceal.
Haematemesis (vomiting blood from the UGI tract), coffee ground vomit, melaena, or rectal bleeding may be present, with or without signs and symptoms of hypovolaemia. The characteristic pungent smell and typical black tarry appearance of melaena may help distinguish it from the sticky but relatively solid, grey-black stools associated with iron therapy.
Rectal bleeding usually indicates a colonic, rectal, or anal lesion, but 15% of patients reporting rectal bleeding have an UGI source.
A history of NSAID use, previous peptic ulcer disease, preceding weight loss, or dysphagia may indicate the cause of bleeding. A history of liver disease or high alcohol intake may suggest the presence of varices, which are associated with severe and life-threatening bleeding.
Several scoring systems have been designed to predict the outcome of UGI bleeds, and the Rockall system remains the most extensively validated.
The Rockall score combines clinical and endoscopic findings, and gives an indication of re-bleeding risk and death following hospital admission for an acute UGI bleed.
The Glasgow-Blatchford score uses blood urea, haemoglobin, shock parameters, presenting features, and the presence of hepatic disease or cardiac failure to predict the need for intervention and is recommended for use at first assessment.
Management of UGI bleeding
Transfer to hospital for oesophagogastroduodenoscopy (OGD) is required to provide diagnosis, prognosis, and treatment. OGD will be offered immediately after resuscitation in unstable patients with severe bleeding and within 24 hours for all others. Bleeding sites can be treated with endoscopic techniques such as adrenaline injection, band ligation, or thermocoagulation.
In non-variceal UGI bleeding, proton pump inhibitors are only offered after OGD, if there are signs of recent bleeding. In variceal bleeding, endoscopic treatment using banding or injection is accompanied by the use of terlipressin and prophylactic antibiotics.
UGI bleeding that is uncontrolled at OGD may require radiological treatment or surgery. Transjugular intrahepatic portosystemic shunt (TIPS) is an interventional radiology procedure that reduces portal hypertension and is used in oesophageal or gastric variceal bleeding. It may be offered to patients with failed endoscopic management of variceal bleeding. The need for surgical management of UGI bleeding is very infrequent.
Lower GI bleeding
Lower GI bleeding (LGIB) is less common than upper GI bleeding but accounts for 15,000 hospital admissions a year. The majority of acute LGIB stops spontaneously and is also less immediately life threatening.
LGIB may be indicated by fresh or altered blood per rectum or melaena. Symptoms such as weight loss, abdominal pain, change in bowel habit, or a past history of diverticular disease may point to a cause.
Diverticular bleeding can be associated with serious LGIB and is often characterised by the absence of abdominal pain.
Anorectal sources of bleeding include haemorrhoids, fistulae, and fissures. These rarely need immediate assessment in hospital unless associated with perianal sepsis. Per rectum examination is essential as it identifies perianal causes and many rectal cancers are palpable.
Unlike UGI bleeding, there are no widely adopted risk scores for LGIB. Empirical evidence supports the use of age, signs of shock, and comorbidities to stratify risk. In addition, the use of NSAIDs significantly increases the risk of serious LGI bleeding. Patients taking antiplatelets, warfarin, or direct oral anticoagulants are also likely to suffer more severe bleeding.
LGIB poses a diagnostic challenge as colonoscopy and flexible sigmoidoscopy offer an opportunity of direct visualisation of the source of the bleeding, but may be limited by poor bowel preparation. Abdominal CT may establish gross causes such as colitis or colorectal cancer, but may miss subtler diagnoses. CT angiography may demonstrate a blush of contrast, but the patient must be bleeding at the time of the scan. In an unstable patient an urgent OGD is prudent to rule out UGI causes.
Most patients do not receive any treatment for LGIB and most commonly the bleeding stops spontaneously. Endoscopic treatment in the form of adrenaline injection or haemoclipping may be offered at colonoscopy or flexible sigmoidoscopy, but much less frequently than UGI bleeding.
Mesenteric angiography may be used to embolise a bleeding vessel, but requires a high volume bleed for adequate visualisation and has a risk of bowel ischaemia. Surgery may be indicated for continued bleeding that does not respond to conservative measures, but this is rare.
- Kathryn Oakland is colorectal research fellow, Oxford University Hospitals
- Rockall TA, Logan RF, Devlin HB et al. Risk assessment after acute upper gastrointestinal haemorrhage. Gut 1996; 38(3): 316-21.
- Button LA, Roberts SE, Evans PA et al. Hospitalized incidence and case fatality for upper gastrointestinal bleeding from 1999 to 2007: a record linkage study. Aliment Pharmacol Ther 2011; 33(1): 64-76.
This is an updated version of an article that was first published in November 2010.