J Royal Naval Medical Service 2014, Vol 100.3 308

The treatment of haematemesis and upper gastrointestinal bleeding in United Kingdom Armed Forces and other deployed units

Surg Lt R Arr Woodward, Surg Cdr M Khan

Abstract

Introduction Upper Gastro-intestinal (UGI) bleeding is a significant cause of morbidity worldwide. United Kingdom Armed Forces (UKAFs) are not immune to this condition. There is a substantial body of conflicting evidence regarding initial management and risk stratification.

Aim To provide the background knowledge and treatment pathways required to assess and manage a patient adequately during the first 24 hours of an episode of UGI bleeding.

Assessment Clinical grading of hypovolaemic is inaccurate, but is a broad indicator of severity; the Rockall Score must not be used to assess requirement for intervention. Where laboratory assets are available, the Blatchford score is adequate to assess requirements for intervention.

Management The principles of hypotensive resuscitation (target systolic blood pressure 90 mmHg for the first hour) hold true for UGI bleeds. In areas where endoscopy is available within four hours, a restrictive pattern of packed Red Blood Cell (pRBC) transfusion may be beneficial. Despite limited evidence of benefit, Proton Pump Inhibitors (PPIs) should be given routinely in UKAFs. Where available, in cases of variceal and non-variceal UGI Haemorrhage without locally available endoscopy, administration of tranexamic acid and somatostatin or octreotide should be considered.

Introduction however, Role One units, in particular Royal Naval or Upper Gastro-intestinal (UGI) bleeding is one of the most light infantry units, often operate outside of effective common causes of admission to hospital in the United local medical care. In these situations, a good knowledge Kingdom (UK), with an annual incidence of 134/100,000 of initial actions, risk assessment and the evidence base (1). This is more common than acute behind further management is essential for Role One (84/100,000 (2)), and with a mortality of approaching 10% medical personnel, as well as those at higher levels of care. it is one of the more significant causes of morbidity in acute settings (1). Specific Challenges for UKAFs UKAFs often operate in austere environments: even at 80% of UGI bleeds are self-limiting (3). The ‘gold standard’ Role Two and Role Three resources are restricted, and the investigation and treatment for all significant UGI bleeds availability of interventions is not guaranteed; this limits is therapeutic Oesophago-Gastro-Duodenoscopy (OGD), the management options available to deployed medical along with fluid and blood product replacement (4). practitioners. United Kingdom Armed Forces (UKAFs) do not deploy with an organic endoscopic capability even at Role Three. Firstly, OGD is not available within operational theatres, so This increases the reliance on local medical facilities; accurate assessment of patients with the aim of identifying 309 Clinical

those most at risk is essential. Secondly, blood products (as a high concentration, and to dissolve in and through well as other products and medications mentioned below) the bicarbonate protective layer of the gastric mucosa. are not always carried and are frequently restricted in Once present in the mucosa, the inhibition of COX 1 quantity even if they are. Therefore, the use of these products mediated prostaglandins rapidly reduces the production of should be rational and limited to those patients most likely bicarbonate, thus rendering the gastric mucosa vulnerable to receive maximal benefit. Finally, although focus is often to erosion. There is little evidence that traditional non- on the acute management of bleeding, Role One also acts pharmacological methods of reducing ulcer formation as a primary care facility for deployed personnel. It is worth (e.g. taking NSAIDs with food) leads to any reduction in bearing in mind that most UGI symptoms or bleeding pathology (9). encountered will be of a chronic nature. H. pylori is endemic throughout the world, and the Epidemiology greatest risk factor for transmission is crowded conditions, The risk factors for UGI bleeding (Box 1) (5,6) suggest that often linked to social deprivation (10). Due to the living UKAFs may be in a higher risk category due to the fact conditions associated with military life, it is reasonable to that alcohol abuse, Non Steroidal Anti-Inflammatory Drug anticipate infection rates to rise with military service or (NSAID) use and smoking are all endemic; this is coupled deployment. An observational study showed that deployed with the psychological stress of operations (5,7). There is US Naval personnel showed an increase in prevalence little data available on the incidence of UGI bleeding in UK following a six-month deployment south of the equator, forces. However, an analysis of United States (US) troops but whether this increase resulted from crowding or from suggests that the incidence is lower in military personnel than contact with local populations was not established (11). others (34/100,000 personnel per year) (7). One theory for this is that age is a major risk factor, and this is a protective A rarer cause of UGI bleeding, in particular profuse factor in a predominantly younger military population. haematemesis, is oesophageal varices. These are swollen veins (varicosities) that arise at the cardiac sphincter, due to Despite this apparent protection, there are specific risk venous diversion secondary to portal venous hypertension. factors on operations to be considered. Trauma is a major The most common cause is liver , and in risk for UGI ulceration with an incidence of 25% in trauma forces and ex-forces populations this is predominately patients requiring Intensive Care Unit (ICU) admission secondary to alcohol abuse. However, in local populations, (5). The mechanism can be direct trauma, ischaemia to the schistosomiasis or chronic viral should be gastric mucosa (Curling’s ulcer), or associated physiological considered. The treatment of variceal haemorrhage is highly stress in intensive care environments. Mortality amongst specialised and will not be covered in detail here. However, in-patients who develop UGI bleeding is higher (8). For the principles and broad management are discussed and this reason, gastric ulcer prophylaxis for trauma patients is remain the same as for non-variceal haemorrhage. now routine. Other causes include Mallory-Weiss tears (tears in the Risk Factors: mucosal lining of the oesophagus due to prolonged retching) • Alcohol Abuse or oesophagitis, as the result of either prolonged retching • Chronic Renal failure or acid reflux. These can be expected in personnel who • Non-Steroidal Anti-Inflammatory Use (NSAID) through intoxication or gastro-oesophageal reflux disease • Age have exposed the oesophagus to either acid or physical • Lower Socio-Economic Class stress. The most severe complication of this is oesophageal • Smoking perforation, which is suggested by severe retrosternal pain • Stress and also surgical emphysema in the throat, or allied chest • Trauma symptoms.

Box 1: Risk factors for UGI bleeds. The common causes of UGI bleeding are listed in Table 1(12). Pathophysiology The most common causes of UGI bleeding and Assessment haematemesis are gastroduodenal erosions and ulceration. As seen in Table 2, significant UGI bleeding rarely presents Erosions in the gastric mucosa are predominately caused with haematemesis in isolation. However, it needs to be by NSAID use or infection by Helicobacter pylori. Non- noted that bright red vomitus (true haematemesis) is a specific Cyclooxygenase (COX)1 and COX 2 inhibitors reliable sign of ongoing UGI bleeding and needs to be dealt (NSAIDs) are weakly acidic and so disassociate freely in with as an emergency. Regardless of the underlying cause the acidic stomach environment, enabling them to retain the initial approach depends on the haemodynamic state of J Royal Naval Medical Service 2014, Vol 100.3 310

Cause Features

Gastric Ulcer Epigastric / right upper quadrant pain Duodenal Ulcer Melaena Coffee ground vomit Erosive Frank haematemesis History of NSAID use Epigastric pain / .

Mallory-Weiss tear Retching Emesis (without blood initially) Coughing Fresh blood streaks in otherwise normal vomitus No melaena or coffee ground vomit

Oesophagitis Reflux Oesophageal Ulcer Dysphagia Nocturnal cough

Malignancy Dysphagia Early satiety Weight loss

Oesophageal Varices Jaundice Portal Hypertensive Gastropathy Ascites Caput medusa Signs of chronic Profuse bleeding Table 1: Causes and features of UGI bleeding. the patient, and a rapid assessment needs to be made. Table young, fit patients (14), especially in active, healthy Armed 2 outlines the basic clinical markers that should be assessed Forces personnel who frequently have resting heart rates (12,13,14). Following assessment, a decision can be made below 60 beats per minute. as to whether urgent volume resuscitation and intervention is required, or if less urgent investigation is appropriate. The gold standard for diagnosis is OGD, as direct

Deficit 0-15% 15-30% 30-40% 40% Pulse 60-90 90-120 >120 >120 Blood Pressure Normal Postural Drop sBP Decreased sBP <90 Pulse Pressure by ~ 30-40 <30 mmHg mmHg Capillary Refill <2 Seconds > 2 Seconds > 3 Seconds > 3 Seconds Respiratory Rate 12-16 16-20 >20 >30 *sBP = Systolic Blood Pressure Table 2: Signs of haemodynamic compromise and the corresponding volaemic state . It is worth noting that signs and symptoms can vary hugely visualisation of a lesion allows for classification as well depending on the patient and his or her physiological reserve. as intervention if required. Since ulcers or other causes of A pulse rate of 80 may represent a relative tachycardia in significance do not show up well on Computer-assisted 311 Clinical

Tomography (CT), there is rarely a requirement for intervention, or at increased risk of death. radiological investigations. Future investigations of UGI The ‘Rockall Score’ has two components: clinical or pre- bleeding may include the use of angiography to identify endoscopic, and full or post-endoscopic scores. It was and treat a brisk bleeding lesion. However, at present, originally derived from an observational study to derive interventional radiology is beyond the capability of any a prognostic scale for mortality. A score of 0-1 equates to deployed unit. a low risk of mortality of <2.4% (13). The Rockall score

Points 0 1 2 3

Age <60 60-79 >80 e r o

c Shock sBP >100 mmHg sBP >100 mmHg sBP <100 mmHg S

l Pulse <100 Pulse >100 a c i n i l C

- Co- Nil Cardiac failure Renal / e r morbidity Major co-morbidity Malignancy P

Diagnosis Mallory-Weiss tear All other Malignancy No lesion diagnosis y p

o Stigmata None or Dark spot only Visible blood c s

o Visible / spurting t d s n o Vessel P E

Table 3: The Rockall scoring system (the full score is both Pre-Clinical and Post Endoscopy components). An erect Chest X-Ray (CXR) is often used to rule out was never intended to provide an estimate of requirement perforation. In cases where there is a high clinical suspicion for intervention. By contrast, the ‘Blatchford Score’ (Table of perforation (tender, rigid abdomen, guarding, absent 4) was designed to provide an indication of both mortality bowel sounds), it should be noted that erect CXR alone is and requirement for intervention (17). An additional risk insufficient to exclude perforation definitively, and further score, the ‘Addenbrooke’s Scale’, has insufficient evidence radiological confirmation should be sought (15). to support its widespread adoption (4, 18).

Venous blood essays are a mainstay of modern assessment. The important clinical question lies in predicting which In the case of UGI haemorrhage their role is both to patients will require intervention. Several studies have enable intervention (haemoglobin levels, clotting screen compared low-risk Rockall scores of 0 to low-risk Blatchford and blood cross match) and for prognostic purposes (urea scores of 0, and all available studies have found the and electrolytes, liver function tests). In particular, urea Blatchford score to be superior (Blatchford sensitivity 98- and electrolytes are essential, because urea levels rise in 100% versus 63-90% for the Rockall score) (4,17,19). Given isolation from creatinine in cases of amino acid catabolism. the requirement of venous blood samples for a Blatchford In the context of UGI bleeding, haemoglobin is broken score, it is not suitable for use forward of Role Two. down into constituent amino acids and digested; the absorption and catabolism of these amino acids provides Given the low sensitivity of Rockall scores, this is not an isolated rise in blood urea levels commensurate with the recommended as a triage tool, limiting its relevance in level of bleeding. This represents an indirect measure of the terms of establishing a prognosis. The Blatchford score severity of bleeding (16). Aside from providing indicators should be used at Role Two or Emergency Department of bleeding, renal or liver dysfunction are both significant (ED) level in order to assess the need for intervention. co-morbidities that may result in a rise in mortality (13). Management Risk stratification Table 6 outlines the stages and current best practice for the Given the variable incidence and mortality, multiple scores initial management and interventions in UGI bleeding and have been developed to identify those patients requiring haematemesis. Fluid resuscitation is a key component of J Royal Naval Medical Service 2014, Vol 100.3 312

Score Value 1 2 3 4 6 Blood Urea > 6.5 > 8 > 10 > 25 Haemoglobin (M) < 13 > 10 < 10 Haemoglobin (F) <12 < 10 sBP 100-109 90-99 < 90 Other Markers Pulse >100 Syncope Melaena Hepatic / Cardiac Failure Table 4: The Blatchford scoring system. management prior to and during definitive management. transfusion and fluid overload (22). This indicates that In recent years, hypovolaemic shock has been managed a restrictive policy is not it itself beneficial, rather that by a strategy of ‘permissive hypotension’, which involves over-transfusion is harmful; this is already accepted titrating intravenous (IV) crystalloids to maintain a mean wisdom. Current UK practice is to offer pRBC transfusion arterial pressure (MAP) of 70 mmHg or a systolic BP of 90 if Hb is under 8 g/dL; there is no evidence provided by mmHg. The theory is that this stabilises any formed clot, this or other studies to change this in UGI Haemorrhage. as well as reducing the dilutional effect of cold, clotting- Transfusion should continue to be offered in exsanguinating factor-poor fluids that may otherwise worsen a consumptive haemorrhage, when HB <8 g/dl, or if there are clinical coagulopathy (20). Previous studies in this area have symptoms or signs of anaemia. focused on trauma or animal models of hypovolaemia; however, there is no obvious reason to differentiate the two Acid suppression has been a long-standing and contentious and some recent studies have treated them as analogous. A issue. The theory is that by maintaining stomach pH at levels 2013 Cochrane Review found no benefit to liberal infusion >6, the clot is allowed to stabilise. In practice, evidence has strategies in any form of hypovolaemic shock and studies failed to produce a clear link between early Proton Pump concurred between trauma and UGI bleed cohorts (21). Inhibitor (PPI) administration and mortality. However, A concern with any permissive hypotension model of successive reviews and RCTs have shown a decrease in resuscitation has been the development of tissue hypoxia, lesions requiring intervention at endoscopy when PPIs lactic acidosis and worsening end organ failure. For this are administered; this is associated with a reduction in the reason, the aim should be to limit the time at hypotensive severity of lesions identified (23). This suggests that high MAP to less than 60 minutes. For this period in the dose IV PPIs would reduce mortality if endoscopy were hypotensive state there is no evidence for worsening tissue not available, as lower levels of invasive therapy would be hypoxia or outcomes (20). required without commensurate increases in re-bleeding levels or mortality. RCT evidence for this is unlikely to Transfusion of packed Red Blood Cells (pRBC) and blood be forthcoming for obvious ethical reasons. Finally, there products (Fresh Frozen Plasma (FFP) or platelets) in is no evidence that PPIs increase mortality if endoscopy hypovolaemic shock is considered current best practice for is available. Taken together, this means that PPIs should hypovolaemic shock as a result of blood loss (20). These be provided in all cases of UGI haemorrhage, either by trials are based on trauma models of hypovolaemic shock deployed UKAFs or where OGD is likely to be delayed. In and UGI bleeds may broadly be considered analogous a secondary care environment with access to OGD urgently (21). Despite this well-validated finding in trauma, there (<24 hours), there is no evidence of benefit or harm. may now be some evidence to suggest that, in non- exsanguinating UGI haemorrhage (<30% of circulating Tranexamic acid, an anti-fibrinolytic agent, is a recent volume), liberal transfusion strategies may be harmful. but well-established adjunct to trauma care and massive A 2013 Randomised Controlled Trial (RCT) showed an haemorrhage, although with limited evidence in UGI increase in survival from 91 % (liberal) to 95% (restrictive) bleed. A 2012 meta-analysis showed a reduced mortality (hazard ratio 0.55; 95% CI 0.33 to 0.92) if pRBC was given in patients with UGI haemorrhage if tranexamic acid was at a Haemoglobin (Hb) threshold of 7 g/dL as opposed to 9 given, compared to conservative treatment (24). This g/dL (22). Whilst interesting, this study has the limitation benefit was annulled if the patients also received timely that emergent endoscopy (under six hours) was available in endoscopy and PPIs. Therefore, in units without access to all centres; this nullifies any conclusions for the majority of OGD, it is indicated, but it is it is unlikely to be useful if UK hospitals or deployed UKAFs, where OGD is routinely there is access to urgent (<24 hr) endoscopy. unavailable within this time scale. On sub-group analysis, the reduction in morbidity and Prokinetic agents such as erythromycin have fallen mortality was achieved by limiting complications of out of use, predominately because there has been no 313 Clinical

demonstrable difference in mortality. Their mechanism of in management of the former, and the use of a balloon action is to remove blood and gastric contents from ulcers, tamponade (via a Sengstaken-Blakemore tube) is an and so they may be of benefit in improving visualisation of accepted, although frequently futile, measure. Meanwhile, bleeding sites, but they are not routinely indicated, and this in all patients with variceal bleeds, prophylactic antibiotics improved visualisation has never translated to a reduction should be routine; their use is well supported by evidence in mortality (25). (26).

Somatastatin (SST) decreases gastric and duodenal pH and Beyond any immediate medical management, OGD is inhibits gastric blood flow, and octreotide (an analogue the gold standard investigation and primary vehicle for of SST) is thought to work in similar ways. A 2006 meta- intervention. Current accepted practice holds that it is analysis concluded that there is good evidence for the use necessary within 24 hours, and that delays of over 24 of somatastatin in non-variceal haemorrhage if OGD is hours result in a rise in mortality from 2.5 to 6.6% (27). delayed, and moderate evidence for the use of octreotide Interestingly, there is no benefit to providing the endoscopy with the same indication (3). For this reason, UKAFs before six hours. When the Blatchford score and mortality should consider using either of these if they are available; data are considered together, it becomes possible to however, at the time of writing, storage and availability predict the requirement for endoscopy, and in patients limit the use of these therapies. with Blatchford scores of 0 it appears acceptable to delay endoscopy (4, 17). In the management of variceal bleeding, two additional Intervention Strategy Fluids (Non-haematogenous) Crystalloid fluids in boluses of 250ml in order to maintain sBP at: 1. 90 mmHg for one hour 2. 100-120 mmHg > 1 Hour Blood products Transfuse pRBC, platelets and FFP in exsanguinating haemorrhage Transfuse pRBC if – Hb < 8 – clinical signs of anaemia Acid suppression High dose PPI’s e.g. Omeprazole 40mg-80mg IV followed by PPI infusion y

p Tranexamic acid If OGD available* – No evidence o c

s If OGD unavailable – Give 15-25mg/Kg o d n

e ProKinetics No evidence of benefit - e r Somatastatin analogous / octreotide If OGD unavailable – Give 50ug IV bolus P Variceal haemorrhage - Give 50ug IV bolus + infusion Endoscopy Blatchford score 0 and stable – delay acceptable Unstable or Blatchford score >1 – urgent OGD H. pylori eradication If tested positive Acid suppression For all patients t s o Anti-platelets Aspirin + PPI if required, P Clopidogrel if high risk (avoid omeprazole) *OGD available indicates access to endoscopy within 24 hours and on site.

Table 5: Interventions in UGI bleeds. components need to be considered. The first is the ability Following endoscopy, the emphasis is on preventing re- to manipulate portal venous pressure and hence bleeding bleeding. Rates of this appear dependent on H. pylori rates, and the second is the tendency for patients with liver status, and colonised patients should undergo eradication failure to decompensate. SST and octreotide are common therapy, as well as medium- to long-term anti-acid J Royal Naval Medical Service 2014, Vol 100.3 314

treatment (e.g. omeprazole 40mg orally once daily) (4). A required for unstable patients or patients requiring common clinical conundrum is the decision whether or not endoscopy. At present, this is through the Royal Air Force to continue anti-platelet agents in the presence of gastric (RAF) aeromedical evacuation chain. It is important to ulcers and cardiovascular disease. The evidence suggests realise that there are additional components to be aware of that the addition of a PPI to low dose aspirin (75mg) all but when undertaking aeromedical evacuation, including the eliminates the risk of re-bleeding (28). A final consideration, need to optimise haemoglobin levels (Hb >10 g/dL) prior if clopidogrel is used, is an inhibition of its antiplatelet to flying. action by omeprazole, but there is no reduction in efficacy with any other PPI (29). The intervention options for UGI Summary bleeding are summarised in Table 5. Figure 1 summarises UGI bleeding is a serious clinical event and in deployed the management of non-variceal UGI bleeding, and UKAFs management options are often limited by lack of incorporates the Blatchford score for risk assessment of all access to OGD. To compensate for this, UKAFs need to bleeding lesions. have an aggressive policy of pre-OGD management; this should include PPIs and tranexamic acid where appropriate It is important to note with the above management pathways and available, as well as somatostatin or octreotide where that at present not all medications are available at all levels available. UGI haemorrhage can result in shock, and the of care. For instance, in the Royal Navy (RN), the Medical principles of managing this are the same as for hypovolaemic Equipment Table for Stores Afloat (METSA) does not shock secondary to trauma; UKAFs should therefore provide tranexamic acid to all afloat platforms. not hesitate to use permissive hypotensive resuscitation strategies. Finally, in less severe cases, the Blatchford score Evacuation to higher level care can be used as a screening tool to accurately predict the Given the limited interventions in deployed environments, need for intervention. evacuation along the casualty retrieval chain will be

References 1. 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;33:66-76. 2. https://online.epocrates.com/dx/indexprint.jsp;jsessionid=9771BC7EEBFE6F2D77F4D61F793079BB?entire=false&iid=290&sid =23 [Accessed 13 Aug 2014] 3. May G, Butler J. The use of vasoconstrictor therapy in non-variceal upper GI bleeds. Emerg Med J 2006;23(9):722-4. 4. Acute Upper Gastrointestinal Bleeding: Management. CG 141: National Institute of Clinical Excellence: London, UK; June 2012. 5. Anwar F, Al-Khayer A, El-Mahrouki H, et al. Gastrointestinal bleeding in spinal injuries patient: is prophylaxis essential? BJMP 2012;6(1):a607. 6. Holster IL, Kuipers EJ. Management of acute nonvariceal upper gastrointestinal bleeding: current policies. World J Gastroenterolog 2013;Mar 21;18(11):1202-7. 7. Yavorski RT, Wong RK, Maydonovitch C, et al. Analysis of 3,294 cases of upper gastrointestinal bleeding in military medical facilities. Am J Gastroenterol 1995;90(4):568–73. 8. Hearnshaw SA, Logan RF, Lowe D, et al. Acute upper gastrointestinal bleeding in the UK: patient characteristics, diagnoses and outcomes in the 2007 UK audit. Gut 2011; 60(10):1327-35. 9. Chan FK, Graham DY. Review article: prevention of non-steroidal anti-inflammatory drug gastrointestinal complications - review and recommendations based on risk assessment. Aliment Pharmacol Ther 2004;May15;19(10):1051-61. 10. Brown LM. Helicobacter pylori: epidemiology and routes of transmission. Epidemiologic Reviews 2000;22(2):283–97. 11. Hyams KC, Taylor DN, Gray GC, et al. The risk of Helicobacter pylori infection among U.S. military personnel deployed outside the United States. Am J Trop Med Hyg 1995;52(1):109-12. 12. Cappel MS, Friedel D. Initial management of acute upper gastrointestinal bleeding: from intial evaluation up to gastrointestinal endoscopy. Med Clin North Am 2008;92:491. 13. Rockall TA, Logan RF, Devlin HB, et al. Risk assessment after acute upper gastrointestinal haemorrhage. Gut 1996;38(3):316-21. 14. Paccagnella RC, Souza JP, Durocher J, et al. A systematic review of the relationship between blood loss and clinical signs. PLoS One 2013;8(3):e57594. 15. Butler J, Martin B. Detection of pneumoperitioneum on erect chest radiograph. Emerg Med J 2002;19:46-47 16. Witting MD, Magder L, Heins AE, et al. ED predictors of upper bleeding in patients without . Am J Emerg Med 2006 May;24(3):280-5. 17. Blatchford O, Murray WR, Blatchford M. A risk score to predict need for treatment for upper-gastrointestinal haemorrhage. Lancet 2000;356(9238):1318-21. 18. Cameron EA, Pratap JN, Sims TJ, et al. Three-year prospective validation of a pre-endoscopic risk stratification in patients with acute upper-gastrointestinal haemorrhage. Eur J Gastroen Hepat 2002;14(5):497-501. 19. Pang SH, Ching JY, Lau JY, et al. Comparing the Blatchford and pre-endoscopic Rockall score in predicting the need for endoscopic therapy in patients with upper GI hemorrhage. Gastrointest Endosc 2010;71(7):1134-40. 315 Clinical

20. American College of Surgeons, Committee on trauma. ATLS: Advanced Trauma Life Support student course manual. 9th ed. Chicago, IL: American College of Surgeons; September 2012. 21. Kwan I, Bunn F, Chinnock P, et al. Timing and volume of fluid administration for patients with bleeding. Cochrane Database Syst Rev 2014;3: CD002245. DOI:10.1002/14651858.CD002245.pub2. 22. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med 2013;368:11- 21 DOI:10.1056/NEJMoa1211801. 23. Sreedharan A, Martin J, Leontiadis GI, et al. Proton pump inhibitor treatment initiated prior to endoscopic diagnosis in upper gastrointestinal bleeding. Cochrane Database Syst Rev 2010;7: CD005415. DOI:10.1002/14651858.CD005415.pub3. 24. Gluud LL, Klingenberg SL, Langholz E. Tranexamic acid for upper gastrointestinal bleeding. Cochrane Database Syst Rev 2012;1:CD006640. 25. Barkun AN, Bardou M, Martel M, et al. Prokinetics in acute upper GI bleeding: a meta-analysis. Gastrointest Endosc 2010;72:1138 26. Jalan R, Hayes PC. United Kingdom guidelines on the management of variceal haemorrhage in cirrhotic patients. Gut 2000;46:iii1- iii15 DOI:10.1136/gut.46.suppl_3.iii1. 27. Wysocki JD, Srivastav S, Winstead NS. A nationwide analysis of risk factors for mortality and time to endoscopy in upper gastro intestinal haemorrhage. Aliment Pharmacol Ther 2012;36:30. 28. Chan FK, Ching JY, Hung LC, et al. Clopidogrel versus aspirin and esomeprazole to prevent recurrent ulcer bleeding. N Engl J Med 2005;352:238–44. 29. Siller-Matula JM, Spiel AO, Lang IM, et al. Effects of pantoprazole and esomeprazole on platelet inhibition by clopidogrel. Am Heart J 2009;157:148.e1–e5.

Authors Surgeon Lieutenant R Arr Woodward MBBS RN General Duties Medical Officer [email protected]

Surgeon Commander M Khan MBBS(Lond) FRCS(GenSurg) AKC RN Consultant Trauma and Upper GI Surgeon General Surgery Lead for Trauma Imperial College Healthcare NHS Trust