The study of
UPPER GASTROINTESTINAL BLEEDING
Dr. Rajesh Kashyap (M.D.)

Part 2.....

REVIEW OF LITERATURE

The first question the clinician must ask when evaluating gastrointestinal haemorrhage is, "Is there G.I. bleeding?" ¹¹

Upper GI bleed is defined as the bleeding from gastrointestinal tract above the ligament of treitz's. It can manifest as hematemesis, melena or both. Hematemesis is bloody vomitus either fresh and bright red or older and coffee ground in character. Melena is shiny black sticky foul smelling stool. It results from degradation of blood and must not be confused with the effects of exogenous stool darkness such as iron and bismuth¹. Melena occurs if enough hemoglobin remains long enough (>14 hrs) in GI tract to be converted to hematin or other hemochromes that blacken stool¹². Severe acute upper GI bleed can cause maroon or bright red stool¹³. Approximately 60ml of blood is required to produce a single black stool, acute blood loss greater than this may produce melena for as long as 7 days. After the stool colour returns normal, tests for occult blood may remain positive for over 9 weeks.³⁷

Historically, information about color and quantity of stool is helpful. Hematemesis, hematochezia and vomitting of "coffee ground" material is evidence of haemorrhage. A positive nasogastric aspirate for red Blood or coffee ground material confirms the presence of upper GI bleed whereas negative aspirate does not rule out an upper GI source of Bleeding. Testing stool or emesis during an acute episode of bleeding is not usually helpful.¹⁴

Historic clues such as hematemesis or melena can help to determine whether bleeding is from upper or the lower GI tract, hematemesis and melena suggest a upper GI bleeding source whereas hematochezia usually suggest bleeding that originate in the lower GI tract. Melena is often caused by upper GI bleeding. The patients who presents with GI haemorrhage even if it appears to be a lower GI bleeding need a diagnostic gastric lavage.¹⁵

"Coffee ground" haematemesis or Melena alone is less predictive of significant haemorrhage than red hematemesis and melena occurring together.¹⁶

Soplepmann J et al in one of the study conducted in Central Finland (CF) and in Tartu county (TC) Estonia found the overall incidence of upper GI bleed was 68.3/100, 000 adults per year in CF and 98.6 in TC. The incidence increased considerably with age : from 3.1 in those aged 20-29 to 314.1 in those more or equal to 80 in CF, and from 13.2 to 299.1 in TC respectively. Incidence rates were twice as low in younger age groups in CF compared to TC, almost equalized in those greater than or equal to 60 and became even higher in those greather than or equal to 80. 60% of the patients (55% men, 79% women) in CF and 49% (35% men 70% women) in TC were greater than or equal to 60. NSAID use before UGI bleed was equally common (46%) in both regions. Peptic ulcer accounted for over 50% of upper gastrointestinal haemorrhage cases both in CF and TC. Mortality rate was 8.1% in CF and 9.9% in TC. The main epidemiological differences between the regions are the lower overall incidence of UGIH, due to the lower incidence of haemorrhage in the younger age groups, and the higher proportion of the elderly patients in CF compared to TC.¹⁷

Blatchford D et al, in a case ascertainment study of 1882 patients aged 15 years and over and treated in hospitals for acute upper GI bleed found an annual incidence of 172 per 100,000 people aged 15 and over. The annual population mortality was 14.0 per 100,000. Both were higher among elderly people and patients resident in areas of greater social deprivation. Overall case fatality was 8.2%. This was higher among those who bled as inpatient after admission for other reason (42%) and those admitted as tertiary referrals (16%). Factors associated with increased case fatality were age, uraemia, pre existing malignancy, hepatic failure, hypotension, cardiac failure and frank hematemesis or history of syncope at presentation.¹⁸

Peptic ulcer remain the commonest cause of upper GI bleeding in nearly all series but there is considerable variation in proportion of patients with variceal bleeding which tends to be low in UK series and high in those from the USA.^19,20

Peptic ulceration accounts for 30-60% of all episodes of UGIB while bleeding is the commonest complications of peptic ulcerations and occurs relatively often, the annual risk being 4%.^21,22

More recent information suggest that the prevalence of peptic ulcer, which often silent, may be higher than previously recognised, at around 20% for patient not taking non-steroidal anti-inflammatory drugs (NSAID) and similar values for patients taking NSAIDS. Several series have suggested that NSAID usage accounts for 22-31% of all ulcer bleeds in elderly patients^23,24.

George F Longstrugh et al in comparison of retrospective and prospective studies of upper GI bleed found that in retrospective studies DU was culprit in 33%, GU-18%, DU and GU 4%, errosive gastritis 13% erosive duodenitis 4%, erosive oesophagitis 8%. Mallory Weiss tear 5% normal 5% miscellaneous 8% and in prospective case series of inpatients DU 32%, GU 28%, DU and GU in 4%, erosive gastritis 8%, erosive duodenitis 2%, erosive gastroduodenitis 2%, erosive oesophagitis 7%, Mallory Weiss tear in 8%, miscellaneous 10% and normal 1%. In retrospective study they also studied the percentage of upper GI bleeding in different age groups, time from onset of bleeding to endoscopy, haemoglobin levels, and mode of presentation. In this study 53% were NSAID user 3% were alcoholic and 22% had major concomitant diseases²⁵.

Prospective studies have shown that up to 90% of patients with cirrhosis will develop esophageal varices²⁶.

The risk of bleeding from esophagogastric varices is 25-35% for both alcoholic and non alcoholic cirrhosis with the majority of initial bleeding episodes occuring within the first year from the time of diagnosis²⁷.

For those patients who survive the initial episodes of bleeding, the risk of recurrent bleeding approaches 70% with most episodes occuring within 6 months of the index bleed²⁸.

Bleeding from oesophagogastric varices accounts for one third of all deaths in patients with cirrhosis and portal hypertension²⁹.

Fresh hematemesis alone, dark hematemesis with melaena, or fresh hematemesis with melena predicted massive bleeding 21%, 24% and 27% respectively³⁰.

Similar results were found in the national American Society of Gastrointestinal Endoscopy (ASGE) Survey of upper GI Bleed with clear nasogastric aspirate and black stool experienced a 5% mortality, those with coffee ground aspirate and black stool an 8% mortality and those with fresh blood in gastric aspirate and black stool 12% mortality³¹.

ASGE report of 1988 states that upper gastrointestinal endoscopy is very accurate means of determining the presence and site of bleeding. Urgent gastrointestinal endoscopy should be considered in all patients with active haemorrhage. Endoscopy should be ideally performed soon after the patients hemodynamic status has been stabilised. If identification of the bleeding source seems appropriate then endoscopy must be done within 24 hours of the bleeding episode³².

A diagnosis with endoscopy will be achieved in 80% of cases with upper acute GI bleed¹³.

In patients with peptic ulcer disease to document visible vessel (bleeding or not bleeding), presence of stigmata of recent haemorrhage. In patients with portal hypertension to know which of the multiple lesions (if present) is bleeding, grading of the esophageal varices, and to know if the oesophageal variceal bleeding has venous spurt or venous ooze, or adherent clot. Rapid Urease Test for H. pylori if there is evidence of ulcer disease.

As the initial steps of management are being carried out, it is important to take a history and perform a physical examination. Although a specific diagnosis can be made this way only occassionally, helpful clues may be obtained. Have there been prior episodes of bleeding? Is there a family history of diseases that cause bleeding? Does the patient have other illnesses that may lead to bleeding, such as cirrhosis, cancer, coagulopathies, connective tissue disorders or amyloidosis ? Has there been prior surgery, such as for peptic ulcer or for the placement of arterial bypass grafts? Does the patient drink alcohal or take nonsteroidal anti-inflammatory drugs (NSAIDs)? Has there been caustic ingestion? Was the bleeding episode preceded by abdominal pain, dyspepsia, or retching? Has the patient had nose bleeds? Examination of the skin is potentially the most helpful aspect of the physical examination. Cutaneous stigmata of cirrhosis, evidence of underlying cancer (acanthosis nigricans, Kaposi's sarcoma) or hereditary vascular anomalies may be found. Findings of pseudoxanthoma elasticum or Ehler-Danlos syndrome may be diagnostic. Further physical findings include lymphadenopathy or abdominal masses (cancer) abdominal tenderness (peptic ulcer) and splenomegaly (cirrhosis or splenic vein thrombosis). Once the patient becomes haemodynamically stable, early endoscopy should be planned to know the site and cause of upper GI bleed¹.

Resuscitations of haemodynamically unstable patients begins in the admitting area with insertion of two large bore venous cannulas. Saline or Ringer's Lactate solution should be infused as rapidly as the patients cardiopulmonary system allows to correct the abnormalities in vital signs. A good rule of thumb is to replace vascular volume as rapidly as it was lost. Thus patients with slow chronic bleeding will require fluid replacement slowly, if at all measurement of Central venous pressure, a pulmonary capillary wedge pressure may be helpful in selected, fragile patients to prevent rapid fluid administration. Supplementation of oxygen may be of benefit in selected cases. Vital signs, urine output and electrocardiogram are monitored frequently. Patients with dynamically significant acute GI bleeding should be admitted to an intensive care unit.

There are no hard and fast rule to transfuse a patient with GI bleeding, patient who continue to bleed despite therapy, who are in shock who have very low hematocrit less than 20 to 25 percent or who have symptoms related to poor tissue oxygenation (eg angina) should be transfused. For asymptomatic patients with hematocrits above 20-25% other factors must be considered. Is hematocrit likely to drop further as vascular repletion occurs? From what hematocrit level could the patient withstand a recurrent episode of bleeding? Is bleeding acute or chronic? What is the likelyhood of recurrent bleeding. If transfusions are deemed unnecessary, iron suppliments should be prescribed after diagnostic tests has been completed. Blood products are transfused to (1) improve systemic oxygen delivery, (2) improve coagulation or both. Patients whose bleeding has ceased require red blood cells. Fresh frozen plasma or platelets should be given only if a patient require more than one circulating volume of blood transfusions (i.e., ten or more units), there is demonstrable deficiency of clotting factors, or severe thrombocytopenia present³³.

The principles of endoscopic therapy of peptic ulcer bleeding is based upon the basis of haemostasis therapy i.e., to thrombose the bleeding artery and different approaches are

THERMAL MODALITIES

* Argon Laser

* Neodymium YAG Laser

* Heater probes

* Electrocoagulation

INJECTION THERAPY

* Adrenaline (vasconstrictor)

* Scleroscent

* Alcohol

Thrombin (Clotting Factor)

OTHER METHODS

* Topical Sprays

* Microwave coagulation

* Mechanical devices

Injection therapy is widely used as first line of treatment for peptic ulcer bleeding clinical trials has shown that it is as effective as other options and that treatment is safe and cheap.³⁵ The possible mechanism of hemostasis is by injection therapy is to induce temponade, vasconstriction or sclerosis of the vessel. Injection of fluid into fibrotic, rigid ulcer causes compression of the bleeding vessel. This hypothesis is supported by the results of clinical trial reported by Lin et al³⁴.

Dilute adrenaline stops active bleeding in a model of Gastrointestinal bleeding and clinical practice whether it causes vasoconstrictive or temponade is still not clear.

Polidocanol, Ethanolamine and sodium tetradecyl sulphate are widely used sclerosants. Most authoritices consider that if sclerosants have a role it is likely to be in prevention of rebleeding rather than treating active bleeding³⁵.

Injection of 98% Ethanol causes profound dehydration of tissues leading to vigorous inflammation, extensive necrosis and ulcer formation but bleeding is rarely stopped by alcohol injection in contrast with efficacy of adrenaline and laser treatment³⁶.

The specific therapy for the treatment of esophageal bleeding are started after the clear understanding that if due to cirrhosis of liver : Prophylaxis for hepatic encephlopathy; to look for sepsis; to assess for severity of cirrhosis.

(1) Endoscopic sclerotherapy (EST) intra variceal injection an endoscopic variceal ligation (EVL); has to be considered at the time of diagnostic endoscopy if rebleeding is there then maximum two sittings of EST with in a week can be considered. If EST/EVL fails than other options (2) Octeriotide or somatostatin infusion has to be considered.

(a) Octeroitide 50-100 µmg IV bolus, followed by an infusion at 25-50 µgm/hr for 48 hrs. (b) Somatostatin 250 µgm/V bolus followed by 250 µgm/hr in infusion for 48 hrs.

(3) Balloon tamponade with sangstaken blackmore tube, initially tamponade can be given by filling balloon with 200ml of air which should be checked radiologically for proper positioning. Maximum duration of tamponade to be given upto 24 hours.⁸

Norman D Grace in his guidelines for diagnosis and treatment of gastrointestinal bleeding secondary to portal hypertension state that sclerotherapy for control of acute variceal bleeding is the currently accepted therapy of choice. If bleeding is not controlled with two sessions, alternative approaches should be sought. Endoscopic Esophageal variceal ligation is a reasonable alternative, Concomitent use of vasoactive drugs (Vasopressin plus nitro glycerine, somatostatin, octeriotide, terlipressin) offers several advantages. They can be started in the emergency room when variceal bleeding is suspected and offers a potentially clearer field for endoscopist, they lower portal pressure. They are the only proven medical therapy for non esophageal sites of bleeding such as portal hypertensive gastropathy a gastric varices more than 2-3 cms below the gastro esophageal junction. For failures of medical therapy the trans jugular intrahepatic Porta systemic Shunt (TIPS) offers a less invasive means of controlling acute bleeding than an emergency surgically created Shunt. In an appropriate Candidates, liver transplantation is an option but is governed by donor availability³⁸.

Endoscopic rubber band ligation of esophageal varices is a technique developed by Stiegman and coworker in an attempt to provide endoscopic treatment for bleeding esophageal varices which is as effective as sclerotherapy, but has a lower incidence of complications.

Over the last decade, sclerotherapy has become the treatment of choice for control of acute variceal bleeding. Controlled trials have shown sclerotherapy to be superior to balloon tamponade, Vasopressin with an initial success rate of 60-90%. Burroughs et al, reported that 62% of the patients had control of variceal bleeding for at least 5 days after a single session of sclerotherapy and 78% were controlled with two sessions. Additional sessions of sclerotherapy were of diminishing value. Three randomized controlled trials 1,2,3 found sclerotherapy and somatostatin or octeriotide infusion to be equally effective with fewer complications in patients treated with somatostatin or octeriotide. Only one study reported a survival benefit for patient receiving treatment with sclerotherapy when compared with balloon tamponade³⁹.

Gimson et al were able to control variceal bleeding in 91% of patients with variceal ligation compared with 92% for sclerotherapy⁴⁰.

Laine et al in an meta-analysis of studies comparing the effect of endoscopic ligation with that of sclerotherapy in the treatment of patients with bleeding esophageal varices found lower rates of rebleeding, mortality complication and also need for fewer endoscopic treatment with ligation as compared to sclerotherapy⁴¹.

Corley et al identified five independent prognostic indicators to identify complication in patients with UGI bleed. These were 1) evidence of portal hypertension as defined by a history of cirrhosis and/or evidence of ascites on physical examination 2) hematemesis 3) red blood on nasogastric tube aspiration 4) initial systolic BP < 100mHg and 5) initial hematocrit < 30%. Presence and absence of these factors help in risk stratification of the patients presenting without compromising patients outcome⁴².

Chung et al conducted a randomised trial comparing Epinephrine (1:10,000 dil) Vs Epinephrine plus absolute alcohol (upto 1ml Aliquots) in 160 patients with active UGI bleeding (spurting or oozing). The definitive hemostasis rate was over 90% for both group and no difference were reported in terms of rebleeding emergency, operation, transfusions given hospital stay and mortality⁴³.

Rauws EAJ et al in their studies stated that the treatment of bleeding peptic ulcer by injection of adrenaline with or without sclerosant (Polidoconol or sodium tetradecyl sulphate) are safe and effective⁴⁴.

Medical therapy may play a role when therapeutic endoscopy is not available. The use of high dose proton pump inhibitors (Omeprazole 40mg bid) reduces the rate of the recurrent bleeding and the need for surgery⁴⁵.

Surgery is considered the therapy of choice for intractable and recurrent bleeding from peptic ulcer disease (PUD). Surgical therapy for PUD should be considered in the following situations; (1) Severe hemorrhage with significant requirement of blood (e.g., >5 units in 24 hrs); (2) Recurrent bleeding after non-surgical management; and (3) difficulty in obtaining sufficient quantities of compatible blood.

The overall assessment of the patient and continued hemodynamic instability are of greater concern than is exceeding an arbitrary number of transfused units.

Arterial vasopressin 0.15-0.20 units per minute, controls bleeding in some patients with stress ulcers, erosive gastritis and peptic ulcer⁴⁶.

Once the bleeding stops ulcer disease management includings H.Pylori therapy with triple therapy can be considered.

The erosive mucosal disease (EMD) can be drug induced or stress induced. In drug induced EMD, the approach is to stop offending drug and measures to neutralise the gastric acidity or to enhance gastric mucosal resistence. If the EMD is due to stress, then management of underlying responsible illness plus measure to neutralise gastric acidity or to enhance gastric mucosal resistance. Duration of therapy is to be decided by the underlying cause responsible for EMD.

(1) Parenteral H₂ receptor antagonist Ranitidine 50mgm IV bolus dose; 12.5mgm/hr IV infusion.

(2) Sucralfate : 1gm dissolved in 30ml saline six hourly.

or surgery has to be considered in certain cases.

Stress ulceration is a cause of GI bleeding in ICU patients especially in those patients who require mechanical ventilation for more than 48 hours or who have coagulopathy⁴⁷.

Prophylactic therapy should be provided for patients at increased risk. Sucralfate, H₂ receptor antagonist, and antacids are all effective in preventing severe bleeding, but sucralfate is associated with a lower incidence of nosocomial pneumonia⁴⁸.

Mallory Weiss tear is a mucosal tear at the gastroesophageal junction. Patient may have history of retching or emesis followed by hematemesis. Most patients with a Mallory Weiss tear stop bleeding spontaneously and seldom have recurrent bleeding. Some patients require therapeutic endoscopic teachniques. H₂ receptor antagonist and proton pump inhibitors are reported to be effective in its treatment⁴⁹.

Tumours, commonly cancer of the stomach accounts for 2-4% cases of upper GI bleed. Lymphoma, sarcoma, leiomyoma may ocassionally present with bleeding^50,51. In one of the study conducted by Savi des TJ et al found tumours in 5% of cases with upper GI bleed. The one year mortality for the whole group was 100%⁵².

All rights reserved with the author, including the rights of reproduction in whole or part, in any form. No part of the publication, images, photographs, design or any other content may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording, electronic replication, or any information storage and retrieval system, without permission in writing from the copyright holder. However the study may be quoted and referred to in any other research, medical or allied work or/and project without any restriction.

Copyright © reserved with Rajesh Kashyap, Year- 2000