Risk assessment

National Institute for Health and Care Excellence (NICE) guidelines recommend that all patients should have a Glasgow Blatchford Score (GBS) calculated pre-endoscopy followed by a full Rockall Score post-endoscopy.5 The Rockall Score was initially created to predict risk of rebleeding and mortality and requires endoscopy for full calculation.9 The GBS can be calculated prior to endoscopy and has been shown to predict the need for intervention (blood transfusion, endotherapy and surgery) or death.10 Although a GBS >12 has been suggested to identify patients who would benefit from early endoscopy, risk assessment scores have not yet been proven to accurately predict patients who need emergency or urgent endoscopy.11

Blood transfusion

Recent studies have supported a restrictive transfusion policy in UGIB aiming to transfuse at a haemoglobin threshold of 7–8 g/dL.6 ,7 ,12 ,13 The recent European Society of Gastrointestinal Endoscopy (ESGE) guidelines now recommend a target haemoglobin between 7 and 9 g/dL, although a higher threshold may be considered in older patients and those with significant comorbidities.7

Timing of endoscopy

The optimal timing of endoscopy remains unclear. NICE, ESGE and BSG variceal haemorrhage guidelines recommend that all patients have an endoscopy performed within 24 hours of presentation and that unstable patients have an endoscopy performed immediately after adequate resuscitation.5–7 A recent Danish study of over 12 000 patients with ulcer bleeding suggested that if circulatory failure or severe comorbidity was present, survival was optimum if endoscopy was undertaken between 6 and 24 hours from presentation. 14

Non-variceal bleeding

The Forrest classification is based on endoscopic appearance of stigmata of recent haemorrhage (SRH) and is used to stratify patients with non-variceal bleeding (NVB) by their risk of rebleeding and mortality (table 1). All patients should have SRH recorded in the index endoscopy report. Patients with evidence of active bleeding (Ia, Ib) or recent bleeding (IIa) should receive endotherapy. In peptic ulcers with adherent clot, there is some debate regarding the optimal management. The ESGE guidelines recommend that these patients should be considered for clot removal then treatment of the underlying lesion.7

There are a variety of endoscopic therapies used to achieve haemostasis in NVB (table 2). Dual modality endotherapy has been shown to be significantly superior to adrenaline monotherapy and is recommended in the management of patients with NVB with SRH.5 ,7

Variceal bleeding

All patients presenting with suspected variceal bleeding (VB) should receive intravenous antibiotics (guided by local resistance patterns) and terlipressin (2 mg intravenous every 4–6 hours unless contraindicated) prior to endoscopy. Variceal band ligation (VBL) is recommended for the treatment of oesophageal VB. This has reduced rebleeding and mortality, and fewer complications compared with sclerotherapy, which should no longer be used.6 The variceal banding device is attached to the distal tip of the endoscope. The bleeding varix is sucked into the transparent cap and a rubber band is deployed ligating the blood flow. Following initial VBL further banding is delivered at 2–4 weekly intervals until complete eradication of the varices is observed, in addition to use of non-selective β-blockers if appropriate.

N-Butyl-2-cyanacrylate injection is the treatment of choice for the treatment of gastric VB, although injection of thrombin may also be considered. N-Butyl-2-cyanacrylate is injected into the bleeding gastric varix and rapidly solidifies occluding the varix.5 ,6 Repeat injections on follow-up may be necessary to complete occlusion of the gastric varices, although the optimal timing of these is unclear.

Topical endoscopic haemostats: Hemospray and EndoClot

Hemospray is a mineral powder delivered via a catheter which is inserted through the endoscope accessory channel. The powder is only effective when delivered to actively bleeding lesions until haemostasis is achieved. Hemospray is currently licensed for use in NVB only. In case studies, it has shown promise as an effective haemostatic modality as a monotherapy, in combination with conventional modalities or as a rescue modality following failure of standard endotherapy. It may also be useful in diffuse areas of bleeding, for example, from a malignant tumour.18 ,19

A recent prospective, observational case series showed that a similar topical agent Endoclot achieved haemostasis in 64% of patients as a primary modality and in 100% of patients following failure of haemostasis with conventional endotherapy.19

Randomised studies are needed to clarify the safety and long-term outcome of these topical endoscopic haemostats to clarify their role should be used in the management of UGIB.

Over the scope clip

The over the scope clip (OTSC) is a new endoscopic device which is attached via a cap to the end of the endoscope in a similar way to a variceal banding device. The bleeding lesion is sucked into the cap similar to an oesophageal varix and the clip deployed. Despite limited data of efficacy, one retrospective study looked at the use of this device in 41 patients with UGIB and reported successful treatment in 85%.20

Doppler endoscopic probe

Successful endoscopic therapy may be confirmed using a Doppler endoscopic probe (DEP). This is inserted via the accessory channel, and a Doppler signal identifies the presence of blood flow in a vessel following treatment of bleeding ulcer. This may reduce the risk of rebleeding by confirming the absence of blood flow; however, a false-negative DEP may occur due to probe position following endotherapy with endoclips or due to vasoconstriction following adrenaline injection. Ulcers with a positive DEP following endotherapy will not always rebleed; therefore, this could lead to over treatment.21

The recent ESGE guidelines recommend the use of hemospray and OTSC if NVB is not controlled using standard therapy but do not currently advocate a role for DEP.7

Hemospray

Hemospray is currently only licensed for use in NVB; however, small cohort studies and case reports have reported the use of hemospray in patients with bleeding from varices and portal hypertensive gastropathy.22–24 Currently, there is insufficient evidence to recommend this in routine practice.

Danis stent

The Danis stent is a metal mesh stent, which is inserted at endoscopy into the lower oesophagus for the treatment of oesophageal varices. Radiological guidance is not required for insertion. It can remain in situ for up to 2 weeks unlike the Sengstaken-Blakemore tube which should be removed after 24 hours. A recent small multicentre randomised control trial comparing the Danis stent with the SBT reported improved bleeding control with the Danis stent, but mortality was similar. 25 Further data on its role in the management of acute VB are required.