Surgery in Practice and Science 1 (2020) 100004

Contents lists available at ScienceDirect

Surgery in Practice and Science

journal homepage: www.elsevier.com/locate/sipas

Review

Acute upper gastrointestinal bleeding: A review

Elroy Patrick Weledji
Department of Surgery, Faculty of Health Sciences, University of Buea, S.W. Region, Cameroon

a r t i c l e i n f o a b s t r a c t

Keywords: Gastrointestinal haemorrhage is potentially life threatening. Resuscitation, clinically diagnosing and treating
Upper gastro-intestinal bleeding the underlying problem are the principles of management. As most gastrointestinal haemorrhage cease spon-
Acute taneously, there is greater importance in non- operative intervention. Radio-embolisation is a growing useful
Resuscitation
adjunct in available armamentarium for non-operative management. The definitive management of upper GI
Endoscopy
hemorrhage is indicated by the overall risk of re-bleeding and morbidity (Rockall risk score). Failure to respond
Radio-embolisation
surgery to endoscopic and medical management is an indication for urgent surgery. New endoscopic techniques and
radiological embolisation have decreased the role of surgery in management, but collaboration between the en-
doscopist and the surgeon remains.

Introduction Aetiology

Gastrointestinal bleeding may occur from any part of the gut and re- The most common sources of upper GI bleeding (location and aspect
mains a significant health problem. Ulcers are the most common cause of the lesions) are summarized in Table 1 [6]. Non-variceal upper GI
of hospitalization for upper gastrointestinal (GI) bleeding. Thus, the vast hemorrhage is the most common complication of peptic ulcers occur-
majority of clinical trials of therapy for upper GI bleeding focus on ulcer ring in 15% of ulcer patients and accounts for the commonest cause of
disease. A population based audit of upper gastrointestinal bleeding in ulcer related deaths [7]. It tends to be more common in patients aged
the Western world gives the incidence of acute bleeding as 103 cases per 60 and older. Oesophagitis is a form of peptic ulcer disease, but usually
100,000 adults per year [1]. Of this, variceal bleeding accounted for only only causes minor acute bleeding. Occasionally, a significant vessel may
4%. Overall mortality from gastrointestinal bleeding in hospital- admit- be involved with consequent massive arterial haemorrhage which must
ted patients was 14% (11% in emergency admissions and 3% among in- be distinguished from variceal bleeding. Carcinoma and lymphoma of
patients). 27% of patients were over the age of 80 years, as compared the stomach, when at an advanced ulcerated stage, commonly bleed
to less than 10% in the earlier studies were mortality was 10–14% [2, but this usually results in occult blood loss, although occasionally will
3]. Thus, both incidence and mortality increased markedly with age. present with acute haemorrhage. The Dieulafoy’s lesion is a ruptured,
The little improvement in mortality over the years despite therapeutic thick-walled artery with little or no associated ulceration. It occurs in
advances must be related to the dramatic shift in the age of the popu- the fundus as a round mucosal defect with a protruding artery at the
lation at risk [1,4]. As seventy to eighty percent of upper gastrointesti- base [2]. Upper gastrointestinal haemorrhage frequently occurs in hos-
nal bleeding in the general population stop spontaneously [1–3], there pital patients being treated for unrelated conditions. Of these, 25% bleed
is greater importance in the simultaneous resuscitation, investigation from acute erosive gastritis and 50% from peptic ulcers [8]. In ulcers in
and active monitoring of the acutely bleeding patient than for operative the posterior wall of the duodenum or the lesser curve of the stomach
intervention. The goal is to stop continuing hemorrhage and decrease the gastroduodenal or left gastric arteries can be respectively involved,
the risk of re-bleed. The formation of specialist gastrointestinal bleed- and these lesions are particularly prone to massive haemorrhage and
ing units with dedicated properly trained endoscopists have minimized re-bleeding after initial stabilization [9]. In addition, the atherosclerotic
the morbidity and mortality of gastrointestinal bleeding [4]. New endo- arteries of the aged will not favour vasospasm/constriction. It is inter-
scopic techniques and interventional radiology have decreased the role esting that peptic ulcer disease may be decreasing but the number of
of surgery in management, but collaboration between the endoscopist operations for bleeding ulcers remain unchanged. This is probably be-
and surgeon is still mandatory. Any trial of new therapeutic techniques cause ∼ 10–20% of peptic ulcers bleed without any antecedent symptoms
must improve upon the natural haemostatic process to be of real benefit. [1], in addition to the increased ingestion of aspirin and non-steroidal
The work has been reported in line with the SCARE 2018 criteria [5]. anti-inflammatory medication in the elderly [4]. Presentations may vary

E-mail address: [email protected]

https://doi.org/10.1016/j.sipas.2020.100004
Received 27 February 2020; Received in revised form 17 April 2020; Accepted 19 April 2020
2666-2620/© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license.
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
E.P. Weledji Surgery in Practice and Science 1 (2020) 100004

Table 1
Differential diagnosis of haemetemesis or melaena

Common Less common (<5%) Rare (1%)

Peptic ulcers/erosions (45%) Duodenitis Hereditary telanngiectasia

Oesophagitis (10%) Oesophageal varices Aorto-duodenal fistula
Mallory-Weiss tear (5%) Gastro-oesophageal cancer Haemostatic defect
Dieulafoy ulcer (5%) Pseudoxanthoma elasticum
Idiopathic (25%) Haemobilia
Pancreatitis
Angiodysplasia
Portal hypertensive gastropathy

Table 2
Rockall risk score

Score
Variable 0 1 3
2
Age <60 years 60–79 years >80 years
Shock ‘No shock’ SBP > 100 mm Hg “Tachycardia” ‘hypotension”
HR < 100 bpm SBP >100 mm Hg SBP < 100 mm Hg
HR > 100 bpm
Comorbidity No major comorbidity CCF, IHD, Major Renal failure, liver failure,
comorbidity disseminated
malignancy
Diagnosis Mallory- Weiss tear, no All other diagnoses
lesion identified and no
SRH
Major stigmata of recent None, or dark spot only Blood in upper GI tract,
haemorrhage (SRH) adherent clot. Visible or
spurting vessel
Age, Shock, Comorbidity = Initial score criteria (maximum additive score prior to diagnosis = 7)
Diagnosis, Major stigmata of recent haemorrhage = Additional criteria (maximum additive score after diagnosis = 11)
SBP, systolic blood pressure; SRH, stigmata of recent haemorrhage.
Higher score, higher morbidity and mortality. If score < 2, risk of rebleeding is 4% and mortality < 0.1%. If score > 5, risk of rebleeding is>24%
and mortality > 11%.

from melaena to occult haemopositive stools, to massive haemeteme- endoscopy, and is predictive of death and re-bleeding in patients with
sis and shock [10]. Initial management of upper GI bleeding involves ulcers or varices. The maximum additive score prior to diagnosis is 7,
resuscitation, followed by endoscopy. Initial resuscitation involves as- and the maximum additive score after after diagnosis is 11. If initial
sessment of airway, breathing, and circulation (ABC), establishing large (pre-endoscopic) score is 0 (age< 60 years, no shock, no co-morbidity),
bore intravenous (I/V) access, urinary catheterization, blood transfu- there is an extremely low risk of death or rebleeding and non-admission
sions as required, correction of clotting abnormalities especially for pa- or early discharge with appropriate outpatient follow-up is considered.
tients with liver disease, and urgent endoscopy if the patient is unsta- If initial (pre-endoscopic) Rockall score is above 0, there is significant
ble. Nasogastric tube insertion and aspiration with or without saline mortality. Score 1 has a predicted mortality 2.4%; score 2 has predicted
lavage of retained blood greatly aids endoscopic visibility. Head down mortality of 5.6% and a score > 8 has a high risk of death. Upper GI
and semi-prone positioning protects against an aspiration pneumonia endoscopy is indicated for full assessment of bleeding risk. If full (post-
which is the commonest cause of death in these patients. If the patient is endoscopic) Rockall score is <3, there is low risk of re-bleeding or death
haemodynamically unstable resuscitation is carried out in the intensive and early discharge and out- patient follow-up is considered [13,14].
care unit, followed by urgent endoscopy or resuscitation in theatre with The mortality predicted by the Rockall score is summarized in Table 2.
urgent endoscopy under general anaesthesia (i.e. the circulation assess-
ment “C” stage of the Advanced Trauma Life Support (ATLS) continues Endoscopic management
in theatre) [10,11]. However, rapid exanguination calls for an imme-
diate operation to stop the bleeding (resuscitative laparotomy) without Endoscopy allows visualization of source of bleeding and therapeutic
prior endoscopy. Emergency endoscopy is both diagnostic and thera- intervention and/or biopsy. Since in the majority of cases bleeding set-
peutic, but the diagnostic yield decrease with time following the initial tles spontaneously, following a pre-endoscopy Rockall score, if the pa-
bleed. Performed early it identifies the cause of bleeding in 90% of cases tient is haemodynamically stable upper GI endoscopy can be done on the
[8–10,12]. The aim is to stop continuing haemorrhage and decrease the next available list. If unstable the best option is endoscopy under gen-
risk of re-bleed [10]. Pre-endoscopic proton pump inhibitor (PPI) may eral anaesthesia in the operating theatre where airway can be protected
be considered to decrease the need for endoscopic therapy but does not to prevent aspiration. Pre-endoscopic erythromycin is considered to in-
improve clinical outcomes [13]. crease diagnostic yield at first endoscopy [13]. An appropriate multi-
channel gastroscope should be available that allows for insertion of de-
Risk stratification of upper GI haemorrhage vices such as needles, probes, clips applicators, whilst allowing irriga-
tion/ aspiration. Endoscopic therapy should only be delivered to actively
The mainstay of investigation and management is endoscopy, and, bleeding lesions, non-bleeding visible vessels, and when technically pos-
the definitive management is indicated by the overall risk of re-bleeding sible to ulcers with an adherent blood clot [13]. The endoscopic stig-
and morbidity [10,12–14]. Re-bleeding can be predicted by the endo- mata of recent haemorrhage can predict the risk of re-bleeding and the
scopic findings and several clinical factors (Table 2). An initial Rock- need for endoscopic haemostasis or surgery (Table 3) [10]. In the best
all scoring system [14] is an appropriate tool for assessment prior to hands, initial haemostasis can be achieved in over 90% of ulcers with ac-
E.P. Weledji Surgery in Practice and Science 1 (2020) 100004

Table 3
Rebleeding rates and need for endoscopic treatment based on endoscopic stigmata of recent haemorrhage (SRH) [10,13].

Endiscopic stigmata Incidence (%) Risk of rebleeding (%) Endoscopic Treatment?

Active arterial bleeding (Forrest 1A, 5–15 90–100 Yes (Endoscopic therapy + IV PPI bonus
1B) and infusion 72 h)
Non-bleeding visible vessel (Forrest 25 50 Yes (Endoscopic therapy + IV PPI bonus
2A) and infusion 72 h)
∗
Adherent clot (Forrest 2B) 15 30 Yes (may consider Endoscopic therapy if
technically feassible + IV PPI bonus and
infusion 72 h)
Flat/red black spots or clean base 15–20 5–10 No (oral PPI)
(Forrest 2C,3B)
No stigmata 35 0–3 No (Oral PPI)
∗
Initial results from a prospective randomized study suggests that where endoscopic expertise is available, firmly adherent
clots could be forcibly removed and the underlying ulcer treated endoscopically.

Table 4
A summary of the endoscopic methods for haemostasis of non-variceal haemorrhage

Method Thermally active methods Injection Mechanical methods Combination methods

Delivery Electrocoagulation Adrenaline: 1:10,000 Endoscopic clips

(Bipolar, Monopolar)
Heater probe Tissue glues Band ligation
Laser Inject + thermal
Purastat haemospray

tive bleeding or visible vessels [10,13,14]. A non-bleeding visible vessel Table 5

(which is really a protruding clot overlying a rent in a non- protruding Predictors of failure of endoscopic treatment
vessel) indicates an early re-bleeding rate of over 50% [10]. It is not
• Haemodynamic instability
known whether endoscopic therapy will be a true substitute for opera- • Significant co-morbidity
tion because visible vessels seen in 45% of cases of ulcer haemorrhage, is • More than 4–6U blood transfusion in 24 h
present in 85% of deaths from that condition [15]. Initial results from a • Endoscopic findings of the ulcer.
prospective randomized study suggests that where endoscopic expertise ○ Actively bleeding vessel
○ Visible vessel
is available, firmly adherent clots on ulcers should be forcibly removed
○ Adherent clot
and the underlying ulcer treated endoscopically [13,16]. None major ○ Ulcer size >2 cm
stigmata of recent haemorrhage such as flat/ red black spots only have
the lowest risk of rebleeding (0–3%) endoscopic treatment is not nec-
essary and oral PPI will suffice (10, 13). Most bleeding duodenal ulcers
can be treated successfully with a combination of endoscopic therapy or non-bleeding visible vessels following endoscopic therapy receive a
(such as adrenalin injection, bipolar diathermy, heat probe, injection high dose intravenous proton pump inhibitor therapy (PPI) therapy (e.g.
sclerotherapy with various sclerosants (alcohol, 1% polidocanol, 3% omeprazole or pantoprazole 80mg bolus followed by 8 mg/h continu-
sodium tetradyl sulphate and 5% ethanolamine), haemoclips/endoclips ous infusion for 72 h). Patients with flat spots or clean-based ulcers do
and laser photocoagulation [15–26]. Definitive haemostasis is higher not require endoscopic therapy or intensive PPI therapy [10, 13,34].
with clipping (86.5%) than injection (75.4%) and clips significantly re- There is insufficient evidence for the use of tranexamic acid or somato-
duce bleeding (9.5%) compared with injection (19.5%) [20,23,26]. Clip- statin in treatment of non-variceal GI bleeding. If patients re-bleed de-
ping and thermocoagulation have comparable efficacy, and there is no spite initial successful endoscopic therapy, they are considered for either
difference in mortality between any intervention [23]. The enhancing repeat endoscopic therapy, selective arterial embolisation or surgery
benefit of combination endoscopic treatment is superior to single modal- [13,34]. About 20–40% require repeat injections for re-bleeding and 5–
ity therapy, and combination treatment does not increase complications 28% require emergency surgery [17,19]. The factors that predict failure
[13,27–29]. The most recommended and popular method is a combina- of endoscopic treatment are haemodynamic instability, significant co-
tion of endoscopic injection of at least 13 ml of 1:10,000 adrenaline morbidity, more than 4–6 U blood transfusion in 24 h and the following
which induces vasospasm, local tamponade and platelet activation plus endoscopic findings of the ulcer: actively bleeding vessel, visible vessel,
either a thermal (bipolar coagulation) or mechanical treatment [17,18]. adherent clot and ulcer size more than 2cm. (Table 5) [10,13,34,35].
New endoscopic techniques include the use of fibrin glue or the simul- These predictors of failure are the indications for surgery (Table 6).The
taneous injection of thrombin and fibrinogen around the base of the approach to a patient with acute upper GI bleed is illustrated in a flow
ulcer and endoscopic ultrasonography-guided angiotherapy (Table 4) chart in Fig. 1. The GI surgeon should be alerted of the possibility of
[15,30]. There are newer products for local haemostasis, e.g. purastat, surgery and should not be far away.
an absorbable hemostatic material and hemospray, a mineral powder
both with very good local effects [31,32]. Endoscopy and endotherapy Radiological embolization
is repeated within 24 h when initial endoscopic treatment is considered
sub-optimal (difficult access, poor visualization, technical difficulties) or Severe bleeding despite conservative medical treatment or endo-
in patients in whom re-bleeding is likely to be life threatening. Patients scopic intervention occurs in 5–10% of patients [4], and requires surgery
with active arterial bleeding or visible vessel who are treated endoscop- or transcatheter arterial embolisation (TAE). Surgery is associated with
ically with successful haemostasis are considered for repeat endoscopy mortality rate as high as 20–40% [36, 37]. Indeed, patients who devel-
and retreatment if necessary after 24 h [13,27,33]. The endoscopic fea- oped failed endoscopic haemostasis are likely to be poor surgical candi-
tures of ulcers direct further management. Patients with active bleeding dates with multiple co-morbidities. Radiological embolisation is highly
E.P. Weledji Surgery in Practice and Science 1 (2020) 100004

Fig. 1. Approach to patient with acute upper GI bleed. ∗ An estimate

of risk can be made before endoscopy (Pre-endoscopy Rock all score).

Table 6 in many institutions as the first- line intervention for massive bleeding
Indications for surgery in non-variceal upper GI haemorrhage from gastroduodenal ulcer after failed endoscopic treatment [36,43].
• Severe haemorrhage not responding to resuscitation
• Recurrence of bleeding after initial control with
endoscopic or medical measures
• Second admission after treatment of ulcer Surgery for bleeding peptic ulcer
haemorrhage
• Prolonged bleeding with loss of 50% or more of blood 10% of patients still require operative treatment to arrest bleed-
volume
ing despite recent advances in medical, endoscopic and interventional
• Blood transfusion more than 4–6U in 24 h
• Age 60 or more with shock or anaemia on admission radio-embolization [44]. Operative intervention is required when bleed-
• Certain endoscopic features of ulcer (ulcer size 2 cm, ing cannot be controlled successfully by endoscopic means or patient is
visible vessel underneath or on base of ulcer, active unstable during initial bleeding requiring a resuscitation laparotomy.
oozing or active arterial bleeding from the ulcer) The indications for surgery are summarized in Table 6 [13]. The princi-
ples of surgery includes an operation that is safest and quickest to arrest
bleeding followed by treatment with PPI and H. pylori eradication ther-
apy [45–47]. General simple suture control of the bleeding combined
with aggressive medical therapy usually proves sufficient. Historically,
useful for patients with recurrent bleeding despite medical or surgi- vagotomy and pyloroplasty with suturing of the bleeding ulcer was the
cal intervention, but local protocols and expertise vary greatly [13,37]. operation of choice but the advent of the aetiological role of H. pylori,
The advantages include the avoidance of surgery, the use in high risk role of NSAIDs, and the development of PPI therapy has made these op-
patients unfit for surgery, advanced malignant gastric ulcer and in oc- erations very rare [48,49]. The special operative hazards would include
cult bleeding not visualized at OGD [38,39]. The bleeding artery is em- (1) the risk of damaging the retroduodenal portion of the bile duct if
bolized by haemostatic substances such as metal coils, oxidized cellu- sutures are hurriedly inserted to underrun a bleeding gastroduodenal
lose, gel foam, or polyvinyl alcohol. Bleeding can also be controlled artery. If a duodenal ulcer has caused much distortion of the tissues the
by selective infusion of vasopressin into the local circulation via the supraduodenal portion of the bile duct is opened and a rubber Jacques
left gastric artery as adjunct for non-operative management of upper GI catheter inserted to aid its identification. The duct is then closed over
bleeding [38]. The use of endovascular embolization is superior with a ‘T’ tube; (2) If no gastroduodenal cause for the bleeding is found, a
high technical (95%) and clinical (72%) success rates [39]. An impor- complete small bowel laparotomy, and careful examination of the pan-
tant complication is the risk of significant ischaemia which increases in creas, gall bladder, bile duct and the aorta where it is crossed by the
patients with previous surgery within the same area [40]. The disadvan- fourth part of the duodenum is necessary. If the site is still not apparent
tage of radiological embolization is that angiography will only detect and the bleeding has stopped, the abdomen is closed. The complications
bleeding vessel if bleeding occurs at a rate of > 0.5 ml/min [38,41]. of surgery would include (i) re-bleeding, which may be from a gastric
The lack of a randomized prospective clinical trial has provided a clin- suture line or from the regional bleeding site. If it persists despite i/V
ical dilemma as to the treatment option between TAE or surgery in re- PPIs and correction of any clotting abnormalities the patient would be
fractory non-variceal upper gastrointestinal bleeding (NVUGIB). A meta- re-operated, the anterior gastric suture line reopened and haemostasis
analysis of studies that directly compared TAE and surgery in haemody- secured; (ii) the discovery of an unsuspected gastric cancer following
namically stable NVUGIB patients following failed endoscopic therapy partial gastrectomy. In a young fit patient, or if the excision margins
showed TAE as safe and effective. TAE had a higher rebleeding rate but appear to be involved with tumour, a more radical elective resection is
did not affect the clinical outcome. It had a slightly lower mortality de- considered; (iii) an incomplete vagotomy, if a definitive truncal vago-
spite the cohort of patients being usually elderly with co-morbidities and tomy and drainage were performed in addition to under-running the
unfit for surgery [42]. This may suggest that TAE is a viable option for ulcer. In this case PPIs should be continued rather than considering an-
the first-line therapy of refractory NVUGIB patients . Another issue to other operation; (iv) leakage from a suture line or duodenal stump if
address is the best treatment option for refractory NVUGIB in haemody- a Billroth II gastrectomy (partial distal gastrectomy with gastrojejunal
namically unstable patients. In the past decade, TAE has been considered anastomosis) was performed [47–49].
E.P. Weledji Surgery in Practice and Science 1 (2020) 100004

Follow-up post discharge since there may be very thickened and haemorrhagic para-oesophageal
tissue and a friable esophagus [59]. In acute bleed, emergency shunt
Prevention of recurrent bleeding is based on the etiology of the surgery has a mortality of 50–80% and thus transection or shunt surgery
bleeding ulcer. Helicobacter pylori is eradicated with triple therapy in- (distal splenorenal shunt or mesocaval shunt) is considered only if phar-
cluding a proton pump inhibitor (PPI) and two antimicrobial agents macologic and endoscopic therapy fails. In preventing bleeding, shunt
(amoxycillin 1 g or clarithromycin 500 mg or metronidazole 500 mg, all surgery has no advantage over sclerotherapy. Urgent liver transplanta-
given bd for 7–14 days, and, after cure is documented anti-ulcer therapy tion for acute variceal bleeding is not practical. The greatest benefit of
is generally not given. Cure of H. pylori should be confirmed 4 weeks af- transplantation is for Child’s Grade C patients who have had success-
ter treatment with C13 urea breath test for duodenal ulcer [50] or repeat ful control of variceal bleeding by the usual treatment measures and
endoscopy for gastric ulcer and if there is suspicion of malignancy. Non- then undergo elective transplantation [64]. An algorithm for the man-
steroidal anti-inflammatory drugs (NSAIDs) are stopped but if they must agement of an acute varceal haemorrhage is shown in Fig. 2.
be resumed low-dose COX-2-selective NSAID plus PPI is used. Patients
with established cardiovascular disease who require aspirin should start Acute erosive gastritis
PPI and generally re-institute aspirin soon after bleeding ceases (within
7 days and ideally 1–3 days). Patients with idiopathic ulcers receive Acute erosive gastritis must be distinguished from the chronic
long-term anti-ulcer therapy [13]. Because of the side effect of platelet forms of gastritis as these do not bleed. Haemorrhagic gastritis is of-
dysfunction with the selective serotonin reuptake inhibitor (SSRI) an- ten caused by stressful stimuli such as head injury (Cushing’s ulcer)
tidepressant, it should be used with caution, or, ideally a non-SSRI may from increased vagal stimulation with resulting acid hypersecretion,
be an appropriate choice in patients who have increased risk of GI bleed- burns (Curling’s ulcer), shock or hepatic failure from probably impaired
ing especially in patients taking NSAID or aspirin [51]. mucosal blood flow. Drugs may also be responsible and the agents
which are commonly implicated include steroids, non-steroidal anti-
Oesophageal varices inflammatory agents (NSAIDS) (from their inhibition of the gastric mu-
cosal protective effect of prostaglandin) and alcohol. Haemorrhage from
Ninety percent of variceal bleeding occurs within 2 cm of the gastro- NSAIDS injury can be arrested with a platelet transfusion but pharma-
oesophageal junction, the site of the porto-systemic venous collaterals. cological suppression of acid secretion with i/v proton pump inhibitors
Evidence of a recent bleed is seen by finding a transparent fibrinous clot (PPI’s) or the use of sucralfate is the mainstay of therapy. A total gas-
on the surface of a varix. Long-term survival is dependent on the severity trectomy is rarely indicated if the site of bleeding is unclear, and it is
of the liver disease (Child’s classification) [52]. The therapeutic arma- important to reduce gastric blood flow as quickly as possible. The first
mentarium include, intravariceal sclerotherapy or endscopic band liga- steps are to ligate and divide the right gastric and gastro-epiploic ves-
tion (for acute bleeding control and future prophylaxis which is effective sels, divide the duodenum, lift up the stomach and ligate and divide the
in most patients); vasopressin/somatostatin(lower portal pressure in the left gastric vessels. [48,60].
acute situation); Sengstaken-Blakemore tube (balloon tamponading of
the varices as a temporizing manoeuvre prior to definitive sclerother-
apy) is hardly ever used anymore; portal/systemic shunts (decrease Mallory–Weiss tear
portal pressure but the progressive increase in hepatic encephalopathy
would render it’s consideration only after two failed sessions of scle- The Mallory–Weiss (M-W) tear occurs in the region of the gastro-
rotherapy); and the transjugular intrahepatic portasystemic anastomosis oesophageal junction on the lesser curve in 80% of cases, as a result of
(TIPS) [53–59]. Oesophageal bleeding stents are increasingly being used severe vomiting or retching, often after excessive alcohol intake. The
[60,61]. 10% of patients will fail initial pharmacologic and endoscopic tear is mostly on the gastric mucosa, but may extend into the oesoph-
therapy. It is important to consider non-oesophageal sites of haemor- agus. Very occasionally repeated vomiting may result in full thickness
rhage, e.g. gastric varices, portal hypertensive gastropathy or ectopic tear, “Boerhave’s syndrome” associated with sudden onset severe up-
varices . Patients with cirrhosis may also bleed from causes unrelated to per abdominal or chest pain. The longitudinal muscle fibres of the oe-
portal hypertension, e.g. duodenal ulcer. The choice of therapeutic op- sophageal wall split above the gastro-oesophageal junction, creating a
tions following failure of standard therapy of oesophageal varices would potential vertical weakness on the left postero-lateral aspect and thus the
include TIPS, surgery or liver transplantation. TIPS is a radiological pro- commonest site of tear in spontaneous oesophageal rupture. Although
cedure which decompresses the portal venous circulation thus lowering bleeding is profuse with M-W, it usually stops spontaneously in 90% of
portal venous pressure, by creating a portal to systemic shunt within cases. Emergency endoscopy establishes the diagnosis and conservative
the liver parenchyma using an expansible stent. Early TIPS is particu- management as with erosive gastritis. Endoscopic electrocoagulation or
larly useful for treatment of gastric varices and considered if pharmaco- operative underrunning is indicated in resistant cases [65].
logic and endoscopy therapy fails. It is effective in controlling bleed 98–
100% of cases with complication rate of 10–25%. Hepatic encephalopa- Aorto-enteric fistula
thy occurs in 15–25% of cases but is severe in only 3-5% [55,58,62].
Early preventive TIPS in patients with oesophageal variceal bleeding Most aortoenteric fistulas are secondary to prior aortic Dacron graft
and decompensated cirrhosis was associated with significant in- hospi- surgery and most always involve the third part of the duodenum [66].
tal reduction in rebleeding and mortality without a significant increase Although a ‘primary’ aorto-enteric fistula can also occur [67], the classi-
in encephalopathy [62]. Numerous operations have been described for cal presentation is a “herald” bleed that occurs and stops spontaneously
portal hypertension when bleeding has not responded to medical treat- hours or occasionally weeks before the major haemorrhage. A high in-
ment, but it is important to remember that some of these operations may dex of suspicion is necessary and the vascular surgeon consulted after
impede a later liver transplant [58]. Oesophageal transection of the oe- upper gastrointestinal endoscopy has excluded other causes of bleeding.
sophageal varices at the gastro-oesophageal junction and re-anastomosis The endoscopist should attempt to reach the third part of the duode-
with a stapler via a small gastrostomy is a relatively simple operation, num to visualize the fistula [68]. An abdominal CT may also demon-
and was found to be more effective than sclerotherapy in controlling strate the fistula. Surgery is required to repair the fistula. This entails
initial haemorrhage in a randomized study but there was no differ- graft removal with either extra-anatomical bypass or in situ replacement
ence in mortality [63]. It can be difficult in those patients who have with a rifampicin-bonded graft [67,69]. A staged endovascular repair
had chronic injection sclerotherapy and contra-indicated when there is without graft removal is palliative treatment as the existing infection of
bleeding from oesophageal ulceration following injection sclerotherapy the graft material inevitably persists and leads to subsequent problems.
E.P. Weledji Surgery in Practice and Science 1 (2020) 100004

Fig. 2. Algorithm for the management of an acute variceal

haemorrhage.

Moreover, endovascular repair following open aortic surgery is techni- Consent

cally demanding given the proximity of the aorto-enteric fistula to the
orifices of the renal arteries [70]. Not applicable.

Conclusions References

As most gastrointestinal haemorrhage cease spontaneously, there is [1] Van Leerdam ME, Vreeburg EM, Rauws EA, et al. Acute upper GI bleeding: did any-
greater importance in non- operative intervention. Radio-embolisation thing change? Time trend analysis of incidence and outcome of acute upper GI bleed-
ing between 1993/1994 and 2000. Am J Gastroenterol 2003;98:1494–9.
is a growing useful adjunct. Although new endoscopic techniques and [2] Rockall TA, Logan RFA, Devlin HB, Northfield TC. Incidence of and mortality from
interventional radiology have decreased the role of surgery, collabora- acute upper gastrointestinal haemorrhage in the United Kingdom. Steering Commit-
tion between the endoscopist and surgeon still remains. Any trial of new tee and members of the National Audit of Acute Upper gastrointestinal haemorrhage.
Br Med J 1995;311:222–6.
therapeutic techniques to stop acute gastrointestinal haemorrhage must
[3] Johnson SJ, Jones PF, Kyle J, Needham CD. Epidemiology and course of gastroin-
improve on the natural haemostatic process, to demonstrate any benefit. testinal haemorrhage in North -East Scotland. Br Med lJ 1973;iii:655–60.
[4] Dronfield MW. Special units for acute upper gastrointestinal bleeding. Br Med J
1987;294:1308–9.
Declaration of Competing Interest
[5] Agha RA, Biowelli MR, Farwana R, Koshy K, Fowler A, Orgill D. For the SCARE
Group, The SCARE 2018 Statement Updating Concensus Surgical Case Report
Nothing to declare. (SCARE) Guidelines. Int J Surg 2018;60:132–6.
[6] Hearnshaw SA, Logan RF, Lowe D, Travis SP, Murphy MF, Palmer KR. Acute upper
gastrointestinal bleeding in the UK: patient characteristics, diagnoses and outcomes
Ethical approval in the 2007 UK audit. Gut 2011;60:1327–35.
[7] Van Leerdam ME. Epidemiology of acute upper gastrointestinal bleeding. Best Pract
Funding: No specific funding was required for this study. Res Clin Gastroenterol 2008;22:209–24.
[8] Schiller KFR, Truelove SC, Williams DG. Haemetemesis and melaena, with special
reference to factors influencing the outcome. Br Med J 1970;ii:7–14.
Author contribution [9] Swain CP, Salmon PR, Northfield TC. Does ulcer position influence presentation or
prognosis of upper gastrointestinal bleeding? Gut 1986;27:A632.
[10] Forrest JAH, Finlayson NDC, Shearmen DJV. Endoscopy in gastrointestinal bleeding.
EPW is the sole author. Lancet 1974;11:394–7.
[11] Anderson I.D. Shock and haemorrhage. Care of the critically ill surgical patient. The
Guarantor Royal College of Surgeons of England; 1999.
[12] Lin HJ, Wang K, Perng CL, et al. Early or delayed endoscopy for patients
with peptic ulcer bleeding. A prospective randomized study. J Clin Gastroenterol
Marcelin Ngowe Ngowe act as guarantor. 1996;22:267–71.
E.P. Weledji Surgery in Practice and Science 1 (2020) 100004

[13] Laine L, Jensen DM. Management of patients with ulcer bleeding. Am J Gastroenterol [41] Poxon VA, Keighley MRB, Dykes PW, et al. Comparison of minimal and conven-
2012;107(3):345–60. tional surgery in patients with bleeding peptic ulcer: a multicentre trial. Br J Surg
[14] Rockall TA, Logan RF, Devlin HB, et al. Risk assessment after acute upper gastroin- 1991;78:1344–5.
testinal haemorrhage. Gut 1996;38(3):316–21. [42] Tarasconi A, Baiocchi G, Pattonieri V, Perrone G, Abongwa HK, Molfino S, et al. Tran-
[15] Jang J-Y. Recent developments in the endoscopic treatment of patients with peptic scatheter arterial embolisation versus surgery for refractory non-variceal upper gas-
ulcer bleeding. Clin Endos 2016;48(5):417–20. trointestinal bleeding: a metaanalysis. World J Emerg Surg 2019;14:3.
[16] Barkun AN, Martel M, Toubouti Y, Rahme E, Bartou M. Endoscopic haemostasis in [43] Loffroy RF, Abualsaud BA, Lin MD, Rao PP. Recent advances in endovascular tech-
peptic ulcer bleeding for patients with high-risk lesions: a series of meta-analyses. niques for management of non-variceal upper gastrointestinal bleeding. WJ Gas-
Gastrointest Endos 2009;69:786–99. trointest Surg 2011;3(7):89–100.
[17] Vergara M, Calvet X, Gisbert JP. Epinephrine injection versus epinephrine injection [44] Hunt PS. Bleeding gastroduodenal ulcers: selection of patients for surgery. World J
and a second endoscopic method in high risk bleeding ulcers. Cochrane Database Surg 1987;11:289–94.
Syst Rev 2007(2):CD005584. [45] Morris DL, Hawker PC, Brearley S, et al. Optimal timing of operation for bleeding
[18] Kim JW, Jang JY, Lee VK, Shim JJ, Chang YW. Comparison of hemostatic forceps peptic ulcer: prsospective randomized trial. Br Med J 1984;288:1277–80.
with soft coagulation versus argon plasma coagulation for bleeding peptic ulcer: a [46] Teenan RP, Murray WR. Late outcome of undersewing alone for gastric ulcer haem-
randomized trial. Endoscopy 2015;47:680–7. orrhage. Br J Surg 1990;77:811–12.
[19] Ginsberg GG, Barkun AN, Bosco JJ, et al. The argon plasma coagulator. Gastrointes [47] Abe N, Takeuchi H, Yanagida O, Sugiyama M, Atomi Y. Surgical indications and
Endosc 2002;55:807–10. procedures for bleeding peptic ulcer. Dig Endosc 2010;22(1):s35–7.
[20] Raju GS, Gajula L. Endoclips for GI endoscopy. Gastrointest Endosc 2004;59:267–79. [48] Steele RJC. The treatment of non-variceal upper gastrointestinal bleeding. In: Grif-
[21] Brock AS, Rockey DC. Mechanical hemostasis techniques in nonvariceal upper gas- fin SM, Raimes SA, editors. Upper gastrointestinal surgery. A companion to specialist
trointestinal bleeding. Gastrointest Endos Clin N Am 2015;25:523–5. surgical practice. WB Saunders Company limited; 2000. p. 361–90.
[22] Kirschniak A, Krutt T, Stuker D, Braun A, Schurr MO, Konigsrainer A. A new en- [49] Weledji EP. Overview of gastric bypass surgery. Int J Surg Open 2016;5:11–19.
doscopic over-the scope clip system for treatment of lesions and bleeding in the GI [50] Jaspersen D, Koerner T, Schorr W, Brennenenstuhl M, Raschka C, Hammar D-H.
tract: first clinical experiences. Gastrointest Endos 2007;66:162–7. Helicobacter pylori: eradication reduces the rate of bleeding in ulcer haemorrhage.
[22] Chan SM, Chiu PW, Teoh AY, Lau JY. Use of the over-the-scope clip for treatment of Gastrointest Endosc 1995;41:5–7.
refractory upper gastrointestinal bleeding: a case series. Endoscopy 2014;46:428–31. [51] Athimulam S, Sharma N, Khan SA. Upper gastrointestinal bleeding in a pa-
[24] Barkun AN, Moodavi S, Martel N. Topical hemostatic agents: a systematic review tient receiving selective serotonin reuptake inhibitor. BMJ Case Rep 2011;2011.
with particular emphasis on endoscopic application in GI bleeding. Gastrointest En- bcr0120113741.
dos 2013;77:692–700. [52] Child CG, Turcotte JG. Surgery and portal hypertension. In: Child CG, editor. The
[25] Sung JJ, Luo D, Wu JC, et al. Early clinical experience of the safety and effectiveness liver and portal hypertension. Philadelphia: Saunders; 1964. p. 50–64.
of Hemospray in achieving hemostasis in patients with acute peptic ulcer bleeding. [53] Burroughs AK, Planas R, Svoboda P. Optimizing emergency care of upper gastroin-
Endoscopy 2011;43:291–5. testinal bleeding in cirrhotic patients. Scand J Gastroenterol Suppl 1998;226:14–24.
[26] Cook DJ, Guyatt GH, Salena BJ, Laine LA. Endoscopic therapy for acute non- [54] D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-ana-
variceal upper gastrointestinal hemorrhage: a meta-analysis. Gastroenterology lytic review. Hepatology 1995;22(1):332–54.
1992;102:139–48. [55] Shiffman ML, Jeffers L, Hoofnagle JH, et al. The role of transjugular intrahepatic
[27] Barkum A, Bardou M, Kuipers EJ, et al. International consensus recommendations portosystemic shunt for therapy of portal hypertension and its complications: a
on the management of patients with nonvariceal upper gastrointestinal bleeding conference sponsored by the national digestive disease advisory board. Hepatology
Upper GI bleeding Consensus Conference Group Consensus recommendations for 1995;22:1591–7.
managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med [56] Yan BM, Lee SS. Emergency management of bleeding oesophageal varices: drugs,
2010;152(2):101–13. bands or sleep? Can J Gastroenterl 2006;20(3):165–70.
[28] Kanwal F, Barkun A, Gralnek IM, et al. Measuring quality of care in patients with [57] Spaander MC, Murad SD, van Buuren HR, Hansen BE, Kuipers EJ, Janssen HL. Endo-
nonvariceal upper gastrointestinal haemorrhage: development of an explicit quality scopic treatment of oesophagogastric variceal bleeding in patients with non-cirrhotic
indicator set. Am J Gastroenterol 2010;105:1710–18. extrahepatic portal vein thrombosis: a long-term follow-up study. Gastrointest En-
[29] Gralnek I.M., Dumonceau J.M., Kuipers E.J., et al. Diagnosis and management of dosc 2008;67(6):821–7.
nonvariceal upper gastrointestinal hemorrhage: European Society of Gastrointestinal [58] Jalan R, Hayes PC. UK guidelines on the management of variceal hemorrhage in
Endoscopy (ESGE) Guideline Endoscopy, 2015;47:a1-a46. cirrhotic patients. Gut 2000;46(suppl 3):1111–15.
[30] Fujii-Lau LL, Wongkee Somg LM, Levy MJ. New technologies and approaches to [59] Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the
endoscopic control of gastrointestinal bleeding. Gastrointest Endosc Clin N Am esophagus for bleeding oesophageal varices. Br J Surg 1973;60(8):646–9.
2015;25:553–67. [60] Khumbhari V, Saxena P, Kashab MA. Self-expandable metallic stents for bleeding
[31] Ibrahim M, El-Mikkary A, Abdel Hamid M, Abdula H, Lemmas A, Mustafa I, De- oesophageal varices. Saudi J Gastroenterology 2013;19(4):141–3.
riere J. Early application of haemostatic powder added to standard management for [61] Rodrigues SG, Cardenas A, Escorsell A, Bosch J. Balloon tamponade and esophageal
oesophagogastric variceal bleeding: a randomized trial. Gut 2019;68(5):844–53. stenting for oesophaeal variceal bleeding in cirrhosis. A systematic review and meta-
[32] Hagel AF, Albrecht H, Nugel A, Vitali F, et al. The application of hemospray analysis. Semin Liver Dis 2019;39(2):178–94.
in gastrointestinal bleeding during emergency endoscopy. Gastroenterol Res Pract [62] Njie B, McCarty TR, Laine L. Early transjugular intrahepatic portosystemic shunt in
2017;2017:308–41. US patients hospitalized with acute esophageal variceal bleeding. J Gastroenterol
[33] Villanueva C, Balanzo J, Torras X, Soriano G, Sainz S, Vilardell F. Value of sec- Hepatol 2017;32(4):852–8.
ond-look endoscopy after injection therapy for bleeding peptic ulcer: a prospective [63] Burroughs AK, Hamilton G, Phillips A, Mezzanote G, McIntyre N, Hobbs KE. A com-
and randomized trial. Gastrointest Endosc 1994;40:34–9. parison of sclerotherapy with staple transaction of the esophagus for the emergency
[34] Barkun A, Sabbah S, Enns R, et al. The Canadian Registry on Nonvariceal Upper Gas- control of bleeding from esophageal varices. NEJM 1989;321(13):857–62.
trointestinal Bleeding and Endoscopy (RUGBE): endoscopic hemostasis and proton [64] Iwatsuki S, Startz TE, Todo S, Gordon R, Tzakis AG, Marsh JW, et al. Liver
pump inhibition are associated with improved outcomes in a real-life setting. Am J transplantation in the treatment of bleeding esophageal varices. Surgery
Gastroenterol 2004;99:1238–46. 1988;104(4):697–705.
[35] Clason AE, Macleod DAD, Elton RA. Clinical factors in the prediction of further [65] Chang MA, Savides TJ. Endoscopic management of non-variceal, non-ulcer upper
haemorrhage or mortality in acute upper gastrointestinal haemorrhage. Br J Surg gastrointestinal bleeding. Gastrointest Endosc Clin N Am 2018;28(3):291–306.
1986;73:985–7. [66] Chopra A, Cieciura L, Modrall JG, Valentine RJ, Chung J. Twenty-year experience
[36] Cheung FK, Lau JY. Management of massive peptic ulcer bleeding. Gastroenterol with aorto- enteric fistula repair: gastrointestinal complications predict mortality. J
Clin N Am 2009;38:231–43. Am Coll Surg 2017;225(1):9–18.
[37] Chiu PW, Ng EK, Wong SK, Teoh AY, Cheung FK, Ying MY, et al. Surgical salvage of [67] Gordon AC, Agarwal M. Primary aorto-enteric fistula. Int J Surg Case Rep
bleeding peptic ulcers after failed therapeutic endoscopy. Dig Surg 2009;26:37–40. 2016;19:60–2.
[38] Loffroy R, Guico B, D’Athis P, Magetta L, Gagnair A, Joine J-L, et al. Arterial emboth- [68] Riddle PJ Jr, Shetty A, Brock A. Confirmatory visualization of an aorto-enteric fistula
erapy for endoscopically unmanageable acute gastroduodenal haemorrhage: predic- with esophagogastroduodenoscopy. Clin Gastroenterl Hepatol 2019;17(13):e157.
tion of early rebleeding. Clin Gastroenterol Hepatol 2009;7:515–23. [69] Kalkos SK, Bicknell CD, Tsolakis IA, Berggvist D. Hellenic Co-operative Group
[39] Van Vught R, Bosscha K, van Minster IP, de Jager CP, Rutten MJ. Embolisation on Aortic Surgery. Management of secondary aorto-enteric and other abdominal
as treatment of choice for bleeding peptic ulcers in high risk patients. Dig Surg artero-enteric fistulas: A review and pooled data analysis. Eur J Vas Endovasc Surg
2007;26:37–41. 2016;52(6):770–86.
[40] Shapiro N, Brandt L, Siprayregar S, Mitsido S, Glotzer P. Duodenal infarction after [70] Lirici MM, tierno SM, Giudce R, Coscarella E, Graziani MG, Pogany G. Secondary aor-
therapeutic gelfoam embolization of a bleeding duodenal ulcer. Gastroenterology to- enteric fistula successfully treated with staged endovascular repair and duodenal
1981;80:176–80. resection without graft removal. Minim Invasive Ther Allied Technol 2019;8:1–6.