Info
🌱 來自: sxtodx
GI bleeding🚧 施工中
GI bleeding
Deepa Rani Nandiwada, MD, MS
CHIEF COMPLAINT
PATIENT
Mr. T is a 66-year-old man who arrives at the emergency department with bloody stools and dizziness. His symptoms started 2 hours ago.
What is the differential diagnosis of GI bleeding? How would you frame the differential?
CONSTRUCTING A DIFFERENTIAL DIAGNOSIS
The approach to GI bleeding is similar to the approach to other potentially life-threatening illnesses. Patient stabilization, specifically, hemodynamic stabilization is the first step in management. In a patient with significant GI bleeding, management precedes diagnosis, usually made by colonoscopy or esophagogastroduodenoscopy (EGD). The pivotal point in the differential, post stabilization, is distinguishing between an upper or lower GI source for blood loss. From there, by using severity of the bleed, patient demographics and assessing risk factors for bleeding sources, one can narrow down the differential further.
Initial management follows a regimented course. First, the patient must have a risk assessment for the severity of bleeding and preparation must be made for resuscitation if further bleeding occurs. This is followed by hemodynamic stabilization and completion of initial diagnostic and therapeutic testing.
A. Risk assessment
1. The best risk assessment tool for upper GI bleed is the Glasgow-Blatchford Score.
a. The score includes variables such as blood urea nitrogen (BUN), hemoglobin, blood pressure, heart rate.
b. Low scores identify low-risk patients who potentially can be treated as an outpatient.
c. A score of 0 has LR– of 0.02 for need for urgent endoscopic intervention equating to a very low-risk patient.
2. For lower GI bleeding, the following patient factors predict poor outcomes (increased risk of mortality, adverse outcomes or re-bleeding):
a. Initial hematocrit < 35%; OR 6.3
b. Age > 60; OR 4.2
c. Gross blood on rectal exam; OR 3.9
d. Heart rate > 100 bpm; OR 3.7
e. Systolic pressure blood pressure < 100 mm Hg; OR 3.0
B. Preparation for hemodynamic stabilization, resuscitation, and further bleeding
1. All patients should have their blood typed and be cross-matched for at least 2 units of packed red blood cells.
2. Two large bore IVs
a. IVs should be ≤ 16 gauge.
b. Because flow = ΔP (πr4/8μL) (where ΔP is the pressure differential, r is the radius of the IV, μ is the viscosity of the fluid, and L is the length of the IV), flow can be maximized by
(1) Increasing the pressure behind the fluid being infused (squeezing the bag).
(2) Decreasing the length of the IV.
(3) Increasing the gauge of the IV (the most effective as the flow goes up by the fourth power of any increase).
c. Large gauge IVs (16 and larger) are much more effective than central lines for volume resuscitation.
Always make sure your patient has 2 usable large bore IVs, so you do not have to worry about IV access should life-threatening bleeding develop.
d. In the setting of severe hemorrhage, a urinary catheter, with regular monitoring of urinary output, helps monitor the adequacy of volume resuscitation.
C. Hemodynamic stabilization
1. Clinically assess volume status.
a. Signs of shock may be seen with 30–40% volume depletion.
b. Orthostasis can be seen with 20–25% volume depletion.
c. Tachycardia may be present with 15% volume depletion.
2. Calculate necessary replacement (weight in kg × 0.6 (lean body weight made up of water) × % volume depletion).
3. Replace fluid losses initially with normal saline or Ringer solution.
4. Consider the need for blood transfusion.
a. There is general agreement on the following recommendations for transfusion in actively bleeding patients.
(1) Patients should receive a blood transfusion when there has been 30% loss of blood volume (manifested by tachycardia, hypotension, tachypnea, decreased urinary output, or CNS symptoms [eg, anxiety/confusion]).
(2) Alternatively, if 2 L of crystalloid have been given without successful resuscitation, blood should be transfused.
(3) If a Hb level is available, actively bleeding patients should receive a transfusion when the level falls below 9 g/dL because it can take time for the Hb to accurately reflect ongoing blood loss.
(4) If large amounts of blood are needed (> 4 units of packed red blood cells), fresh frozen plasma and platelets should also be given.
b. In patients who are not bleeding, withholding transfusions until the Hb is < 7 g/dL is appropriate.
(1) This recommendation is based mainly on a study that demonstrated a mortality benefit associated with a restrictive transfusion strategy.
(2) The exclusion criteria in this study were massive exsanguinating bleeding, an acute coronary syndrome, symptomatic peripheral vasculopathy, stroke, or transient ischemic attack.
(3) All patients had endoscopy within 6 hours of presentation.
(4) Patients in this study also received blood if symptoms of anemia or massive bleeding developed or if they required surgery.
c. In patients with cardiovascular disease, a transfusion threshold of 8 g/dL is probably appropriate.
d. It is important to remember that patients may initially have a normal hemoglobin level when they present with an acute hemorrhage. Hemoglobin level will only fall after fluid resuscitation.
Even after a large hemorrhage, patients may initially have a normal Hb level. The level will only fall after fluid resuscitation.
5. Review the patient’s medication list for antiplatelet and anticoagulant medications. The decision to reverse an anticoagulant is based on a careful balancing of the severity of bleeding versus the need for anticoagulation.
6. If there is concern for an upper GI bleed, empirically start an IV proton pump inhibitor or octreotide or both based on risk factors for peptic ulcer disease and variceal bleeding.
a. IV proton pump inhibitors help reduce the number of endoscopic interventions and risk of re-bleeds in peptic ulcer disease without a mortality benefit.
b. Intermittent IV or oral bolus dosing of proton pump inhibitors is probably equivalent, if not superior to, continuously delivered proton pump inhibitor.
D. Initial diagnostic tests
1. CBC and platelet count
2. Basic metabolic panel (chem-7)
3. Liver biochemical tests (Abnormal results raise the risk of underlying severe liver disease and thus coagulopathy and varices.)
4. Prothrombin time and partial thromboplastin time
5. Upright chest radiograph
a. Most important if there is abdominal tenderness to assess for free air in the abdomen arising from a perforated viscus
b. Patients taking immunosuppressants (including corticosteroids) may have free air in the abdomen with only mild abdominal symptoms.
c. Radiographs may, infrequently, provide other diagnostic clues.
6. Nasogastric (NG) tube placement, which was once considered standard, is now controversial.
a. An NG tube is a minimally invasive way to assess the acuity of bleeding and to help localize its source.
b. Blood or coffee grounds in an NG aspirate is indicative of an upper GI bleed (LR+, 9.6).
c. On the other hand, a negative lavage does not exclude an upper GI source.
d. Research has suggested that NG tube placement does not change patient outcomes (mortality, length of stay, volume of transfused blood, need for surgery).
The differential diagnosis of GI bleeding is based on an anatomic framework. Upper GI bleeds originate proximal to the ligament of Treitz, while lower GI bleeds originate from areas distal to the ligament and are primarily colonic. Therefore, the pivotal points in a patient’s evaluation are characteristics that argue for an upper or lower GI source. Among others, these include history of previous upper or lower bleeding episodes, epigastric pain, melena, and blood in an NG lavage, all of which suggest an upper GI bleed. The causes of upper and lower GI bleeding are arranged below in the approximate order of frequency. Bleeding from a small bowel source is less common. The last category is anorectal bleeding. These are generally smaller bleeds with limited potential to cause hemodynamic instability.
A. Upper GI bleeds
1. Common
a. Peptic ulcer disease
b. Esophageal or gastric varices
c. Mallory-Weiss tear
2. Less common
a. Angiodysplasia
b. Gastritis
c. Malignancy
d. Esophagitis
e. Dieulafoy lesion
B. Lower GI bleeds
1. Common
a. Diverticulosis
b. Malignancy or polyp
c. Colitis
(1) Inflammatory
(2) Infectious
(3) Ischemic
d. Colonic angiodysplasia
2. Less common small bowel sources
a. Angiodysplasia
b. Ulcers
c. Malignancy
d. Crohn disease
e. Meckel diverticulum
C. Anorectal bleeding
1. Hemorrhoids
2. Anal fissures
Mr. T was well until this morning. Abdominal cramping developed while he was eating breakfast. He did not have nausea. He went to the bathroom and passed a large bowel movement of stool mixed with blood. Afterward, he felt better and went to lie down. About 30 minutes later, he had the same sensation and this time passed what he described as “about a pint” of bright red blood. While getting up from the toilet, he became dizzy and had to sit on the bathroom floor for 15 minutes before he could crawl to the phone to dial 911.
At this point, what is the leading hypothesis, what are the active alternatives, and is there a must not miss diagnosis? Given this differential diagnosis, what tests should be ordered?
RANKING THE DIFFERENTIAL DIAGNOSIS
The lack of nausea, vomiting, or abdominal pain, and the presence of bright red blood per rectum are pivotal points in this case and make a lower GI source most likely. Cramping is often seen with GI bleeds, caused by blood passing through the bowel. The patient’s age, acuity, and volume of blood makes bleeding from diverticuli, colitis, malignancy, or angiodysplasia the most likely diagnoses. Whether he has had a recent change in bowel habits, weight loss, or previous bloody stools is unknown; all these factors would heighten suspicion for colitis or malignancy. The amount of blood loss and symptomatic dizziness make hemorrhoids or fissures unlikely. Upper sources of bleeding must also be considered. A brisk bleed from an upper source can present with bright red blood per rectum. Assuming there is no history of liver disease, peptic ulcer disease would be the most likely upper source. Table 19-1 lists the differential diagnosis.
Table 19-1. Diagnostic hypotheses for Mr. T.
Blood is a cathartic. A brisk bleed from an upper source can present with bright red blood per rectum.
Mr. T reports no recent illness or change in bowel habits. He reports no family history of colon cancer, and he has never had a colonoscopy. He has a fifty-pack year smoking history and quit about 6 years ago. He reports drinking 2–4 beers each night.
On physical exam, Mr. T looks anxious but is otherwise well. While sitting, his BP is 120/92 mm Hg and his pulse is 100 bpm. While standing, his BP is 100/80 mm Hg and his pulse is 122 bpm. His temperature is 37.0°C and his RR is 16 breaths per minute. There is no conjunctival pallor. Lungs and heart exams are normal. There are hyperactive bowel sounds but the abdomen is soft, nontender, and with no organomegaly. Rectal exam reveals bright red blood.
Is the clinical information sufficient to make a diagnosis? If not, what other information do you need?
Leading Hypothesis: Diverticular Bleed
Textbook Presentation
The typical presentation is an episode of bright red blood per rectum in an older patient. There may be abdominal cramping but no real pain. A history may include previously diagnosed diverticuli (on a screening colonoscopy, for instance) and possibly a previous, self-limited hemorrhage.
Disease Highlights
A. Diverticular bleeds are the most common cause of lower GI bleeding.
1. The prevalence of the various causes of GI bleeding varies from study to study.
2. One large review gave the following data:
a. Diverticulosis: 35%
b. Inflammatory bowel disease (IBD) or other colitis: 14%
c. Colonic malignancy or polyp: 7%
d. Angiodysplasia: 3%
e. Anorectal cause: 12%
B. The risk of diverticular hemorrhage in a patient with diverticuli is not known but is estimated to be 3–15%.
C. Data from case-control studies suggest that nonsteroidal anti-inflammatory drug (NSAID) use and hypertension are risk factors for diverticular hemorrhage.
D. Although diverticula are most commonly left sided, right-sided lesions are responsible for most bleeding episodes.
E. Bleeding occurs as a vessel is stretched over the dome of a diverticulum. Luminal trauma likely leads to bleeding from the weakened vessel.
F. Spontaneous cessation and only moderate blood loss is the rule, but recurrence is common.
1. About 75% of patients experience spontaneous cessation of hemorrhage.
2. Nearly all patients require < 4 units of packed red blood cells.
3. Approximately 40% of patients have recurrent bleeding.
G. Diverticular hemorrhage carries a poor short-term prognosis.
1. In general, lower GI bleeding carries a better overall prognosis than upper GI bleeding with about half the mortality rate.
2. Mortality rates for diverticular hemorrhage are higher (11% at 1 year and 20% at 4 years) although the cause of death is rarely related to the GI hemorrhage.
Although diverticular hemorrhage seldom causes death, it is a marker for a relatively poor, short-term prognosis.
Evidence-Based Diagnosis
A. History and physical exam
1. The first step in making the diagnosis of any GI bleed is to determine whether the source of the bleeding is the upper or lower tract. Other than frank hematemesis, only a few features are strongly predictive in localizing the site of bleeding to the upper or lower tract. These are outlined in Table 19-2.
Table 19-2. Test characteristics of clinical findings for upper vs lower GI bleeding.
a. Patients who are volume depleted, orthostatic, or hypotensive are about twice as likely to have an upper GI bleed than a lower GI bleed.
b. Because urea is produced as blood is processed throughout the GI tract, a BUN/creatinine ratio > 30 suggests an upper GI source (sensitivity, 39%; specificity, 94%; LR+, 7.5; LR−, 0.64**)**
c. Although hematochezia generally suggests a lower GI source of bleeding, 10–15% of patients with hematochezia have an upper GI source. These patients are more likely to be older and to have duodenal ulcers.
10–15% of patients with hematochezia have an upper GI source of bleeding.
2. Beyond differentiating between upper and lower sources of GI bleeding, certain historical features may point to a specific diagnosis (Table 19-3).
Table 19-3. Historical features in the diagnosis of GI bleeding.
a. These features should be sought in every patient with GI bleeding.
b. They are, however, only suggestive and by no means diagnostic.
B. Endoscopy
1. In a patient with GI bleeding, EGD is usually recommended as the first procedure unless the suspicion for a lower GI bleed is very high (based on history and, possibly, a negative NG tube aspirate). This recommendation is based partly on the higher potential for life-threatening blood loss from upper GI bleeds.
2. Colonoscopy
a. The diagnosis of diverticular hemorrhage is usually made on colonoscopy.
b. It is important to realize that this diagnosis is usually presumptive (87% of the time in some studies) based on seeing diverticula and blood in the same region of the colon.
c. Less commonly, a definitive diagnosis is made when active bleeding or stigmata of recent bleeding in a diverticulum is seen.
C. Radionuclide scintigraphy
1. Uses either radiolabeled sulfur colloid or labeled red blood cells.
2. Can detect bleeding as slow as 0.05 mL/min.
3. Most commonly used for detecting the source of bleeding in patients with persistent bleeding and normal upper and lower endoscopies.
4. Also used to localize bleeding prior to resection or angiography.
5. The test characteristics of radionuclide scintigraphy in GI bleeding are not well defined, varying from study to study.
a. In a representative study, only 39% of patients had positive scans (sensitivity = 39%).
b. In this study of patients who had further evaluation of their bleeding, 48% were found to have bleeding at the sight of the positive scan and 10% were found to have bleeding at a different site.
c. Scans in patients who recently required transfusion are most likely to be positive; scans that turn positive quickly are best at localizing bleeding (about 95% accurate).
D. Angiography
1. Requires bleeding at a rate of about 0.5 mL/min to detect active bleeding.
2. Sensitivity is about 50% (though that number depends greatly on selection of patients).
3. Like radionuclide scintigraphy, angiography can be useful for localizing the site of bleeding before surgery and is considered more reliable.
4. An angiographic procedure also presents the option of catheter-based intervention.
Treatment
A. Management of blood loss
1. As discussed above, all GI bleeds call for similar treatment of a patient who has lost, or has the potential to lose, a significant amount of blood.
2. Patients need to be closely monitored for signs of bleeding (increasing tachycardia, orthostasis, oliguria, declining Hb).
3. Typically, patients have a CBC checked every 6 hours until stability has been achieved but the intensity of monitoring varies with risk of rebleeding.
B. Management of diverticular hemorrhage
1. Because most diverticular hemorrhages stop spontaneously, specific treatment is often not necessary.
2. Endoscopic treatment is primarily clipping, although thermocoagulation and sclerotherapy are occasionally used.
3. Angiographic intervention, with vasoconstrictor agents or embolization, can also be used. Occasionally, local vasopressin infusion may be a temporizing measure.
C. Colectomy
1. Curative therapy for diverticular bleeding is removal of the portion of the colon containing the diverticula.
2. Recommended for either persistent, large bleeds (over 4 units in 24 hours or 10 units during the course of a single bleed) or for frequent recurrences.
The diagnosis of diverticular hemorrhage is often presumptive. Localization of the bleeding site before surgery must be as definitive as possible.
MAKING A DIAGNOSIS
Mr. T was given 1 L of normal saline. While in the emergency department, he again passed a large amount of bright red blood.
Initial laboratory tests are normal. Important values are BUN, 12 mg/dL; creatinine, 1.1 mg/dL; Hb, 13.9 g/dL. The patient was admitted to the medical ICU.
Have you crossed a diagnostic threshold for the leading hypothesis, diverticular bleed? Have you ruled out the active alternatives? Do other tests need to be done to exclude the alternative diagnoses?
Mr. T weighs 75 kg and his orthostasis suggests 20% volume depletion. His fluid deficit is about 9 L (75 kg × 20% volume depletion × 60%). Assuming this deficit is all from the GI bleed, it is very likely that his Hb will fall once he is hydrated.
His hematochezia and normal BUN/creatinine ratio are suggestive of a lower GI bleed. Although Mr. T is relatively young and does not have any comorbidities, he was admitted to the ICU because he has signs of hemodynamic instability, including orthostasis with tachycardia, and has shown evidence of active bleeding. Hemodynamic instability and active bleeding each independently are associated with tripling the risk of poor outcomes, such as rebleeding and mortality. Following stabilization, initial endoscopy with either colonoscopy or EGD would be reasonable.
Alternative Diagnosis: Angiodysplasia
Textbook Presentation
Bleeding from angiodysplasia can look like any other cause of lower GI bleeding. It is seen almost exclusively in older adults and can present with anything from hematochezia to occult blood loss. In general, hemorrhage from angiodysplasia tends to be less brisk than bleeding from diverticula.
Disease Highlights
A. Angiodysplasias, also called arteriovenous malformations, are dilated submucosal veins that are most commonly seen in the right colon of adults over age 60.
B. Present in < 5% of patients over age 60.
C. Most patients with angiodysplasias do not bleed and those that do tend to have occult blood loss rather than brisk, overt hemorrhage.
D. Angiodysplasia has historically been associated with various diseases (eg, aortic stenosis, cirrhosis) but only a relationship to end-stage renal disease seems definite.
Evidence-Based Diagnosis
A. As in diverticular hemorrhage, colonoscopy, tagged red blood cell scan, and angiography are all used in the diagnostic evaluation.
B. Colonoscopy is the most common tool. It allows good visualization of the cecum, which is the site of most angiodysplasias.
C. If suspicious vascular patterns are seen during colonoscopy, angiography can provide evidence of a diagnosis even without active bleeding.
D. As in diverticular hemorrhage, the diagnosis is often presumptive, made on the basis of visualizing nonbleeding angiodysplasia in a patient with GI bleeding.
Treatment
A. Both acute and chronic bleeding is generally treated endoscopically with thermal or laser ablation. This method can be repeated for recurrent bleeding.
B. Angiographic intervention, with vasoconstrictor agents or embolization, is rarely used.
C. Surgical management (right hemicolectomy) is sometimes required for frequent, recurrent bleeding.
D. Hormonal therapy with estrogen has been used to prevent recurrent bleeding in angiodysplasia, but recent studies suggest that this is not very effective.
E. Whenever possible, long-term antiplatelet therapy should be discontinued.
Alternative Diagnosis: Colon Cancer
Colon cancer is discussed in Chapter 2, Screening & Health Maintenance.
CASE RESOLUTION
Six hours and 3 L of normal saline after his initial Hb of 13.9 g/dL, a repeat Hb was 10.3 g/dL. Given the clinical suspicion of a lower GI bleed, colonoscopy was done about 6 hours after admission. There were multiple left-sided diverticula and a right-sided diverticulum with a nonbleeding visible vessel. A diagnosis of a diverticular hemorrhage was made.
Mr. T remained clinically euvolemic and his Hb stabilized around 10.0 g/dL. He remained in the hospital for 48 hours during which there was no recurrent bleeding and his Hb remained stable. No further treatment (eg, surgery) was necessary.
CHIEF COMPLAINT
PATIENT
Mr. M is a 39-year-old man who arrives at the emergency department after vomiting blood. He reports waking the morning of admission with an “upset stomach.” He initially attributed this to a hangover. After about an hour he vomited “a gallon of blood” with no other stomach contents. Almost immediately afterward, he had a second episode of hematemesis and called 911.
At this point, what is the leading hypothesis, what are the active alternatives, and is there a must not miss diagnosis? Given this differential diagnosis, what tests should be ordered?
RANKING THE DIFFERENTIAL DIAGNOSIS
Mr. M is having an upper GI bleed. The hematemesis is a pivotal point in this case and localizes the source of the bleeding to above the ligament of Treitz. Peptic ulcer disease and gastritis are the most common causes of upper GI bleeding. Although not always present, preceding symptoms of abdominal distress are common with peptic ulcer disease and gastritis. When incorporating patient-specific risk factors, such as this patient’s alcohol use, esophageal varices should be added to the differential diagnosis. The details of the patient’s alcohol use are still unknown, so his risk of portal hypertension cannot be predicted. A Mallory-Weiss tear is also possible, but the patient would report vomiting before the onset of bleeding. Table 19-4 lists the differential diagnosis.
Table 19-4. Diagnostic hypotheses for Mr. M.
On further history, the patient reports no previous episodes of GI bleeding. He reports occasional stomach upset, usually following drinking binges. He denies NSAID use. Mr. M says that he has been drinking heavily since his late teens. He drinks at least a fifth of hard liquor and a 6-pack of beer daily for the last 20 years. He reports that he has not seen a doctor since his pediatrician.
On physical exam, Mr. M is anxious and appears tired. He smells of alcohol. While sitting, his BP is 140/80 mm Hg and his pulse is 100 bpm. While standing, his BP is 100/80 mm Hg and his pulse is 130 bpm. His temperature is 37.0°C and RR is 16 breaths per minute. Sclera are slightly icteric. Lungs are clear and heart is tachycardic but regular. Abdomen is soft without hepatomegaly. There is no ascites, but the spleen is palpable about 2 cm below the costal margin.
Given the alcohol history, scleral icterus, and splenomegaly, a hemorrhage from esophageal varices needs to move above peptic ulcer disease on the differential diagnosis.
Is the clinical information sufficient to make a diagnosis? If not, what other information do you need?
Leading Hypothesis: Esophageal Variceal Hemorrhage
Textbook Presentation
A patient with known cirrhosis presents with heavy upper GI bleeding (hematemesis or melena). Stigmata of chronic liver disease as well as a history of previous hemorrhages are frequently present. Laboratory data demonstrate liver biochemical tests consistent with cirrhosis and thrombocytopenia.
Disease Highlights
A. Esophageal varices are portosystemic collaterals that dilate when portal pressures exceed 12 mm Hg.
B. Although varices are the second most common cause of upper GI bleeding, they account for 80–90% of GI bleeds in patients with cirrhosis.
C. Gastroesophageal varices are present in about 50% of patients with cirrhosis.
D. The prevalence of varices depends on the severity of the cirrhosis.
E. The Child-Pugh system classifies patients based on the severity of their cirrhosis.
1. The system takes into account the presence of encephalopathy, ascites, hyperbilirubinemia, hypoalbuminemia, and clotting deficiencies (Table 19-5).
Table 19-5. Child-Pugh classification.
2. 40% of patients with Child-Pugh grade A disease have varices, while 85% of patients with grade C disease have varices.
F. Approximately 33% of patients with varices experience hemorrhage.
G. Varices may develop from cirrhosis of any cause.
H. Of all GI bleeds, those from varices carry the worst prognosis.
1. Nearly 33% of patients die at the time of their first variceal hemorrhage.
2. Up to 60% of survivors have recurrent bleeding in the first year.
3. A variceal bleed carries a 32–80% 1-year mortality and a 6-week mortality of 15–20%.
4. A hepatic venous pressure gradient > 20 mm Hg predicts poor outcomes.
Esophageal varices are, by far, the most lethal type of GI bleeding.
Evidence-Based Diagnosis
A. Esophageal varices are diagnosed with endoscopy.
B. Screening for varices
1. Because variceal bleeding carries such a high mortality, the goal is to detect varices before they bleed so that prophylactic treatment can be initiated.
2. All patients with cirrhosis should undergo screening endoscopy every other year.
3. Patients with cirrhosis but without splenomegaly or thrombocytopenia are at the lowest risk for having varices (about 4%). Endoscopy may be delayed in these patients.
C. Of all causes of GI bleeding, varices are probably the easiest to predict. One study has the sensitivity and specificity of physicians predicting variceal hemorrhage at 82% and 96%, respectively.
Treatment
A. Primary prophylaxis (patients with varices but no previous bleeding)
1. Nonselective beta-blockers (usually propranolol or nadolol) effectively decrease portal pressures.
2. Patients at higher risk for bleeding should also undergo band ligation of the varices.
B. Secondary prophylaxis (patients who have had a previous variceal hemorrhage)
1. Portosystemic shunt procedures, either surgical or transjugular, should be considered.
2. Liver transplantation is the definitive therapy.
C. Treatment of acute hemorrhage
1. Even more than other GI bleeds, achievement of hemodynamic stability in variceal bleeds is of primary importance because the hemorrhage is potentially massive.
2. Patients with variceal bleeding are at high risk for bacterial infections, especially spontaneous bacterial peritonitis. Administration of antibiotics (ceftriaxone or norfloxacin), given prior to endoscopy, have been shown to decrease both the rate of bacterial infections and mortality.
3. Attention should be paid to the patient’s coagulation status.
a. Given the prevalence of liver disease in these patients, there is often coagulopathy related to factor deficiency or thrombocytopenia.
b. Platelet transfusions are recommended for levels under 50,000/mcL during an active bleed.
4. Somatostatin or octreotide should be given as soon as variceal hemorrhage is suspected. These drugs decrease portal pressure and decrease bleeding.
5. Endoscopic banding or sclerotherapy are done initially and if bleeding recurs.
6. Balloon tamponade may be used as a temporizing measure prior to endoscopic therapy.
7. Transvenous intrahepatic portosystemic shunting (TIPS) is an option for patients who continue to bleed despite intervention. In high-risk patients, such as Child-Pugh class C, early shunt placement has a survival benefit.
8. Surgical intervention is seldom called for as the mortality is extremely high.
MAKING A DIAGNOSIS
NG tube lavage in the emergency department revealed bright red blood that did not clear with flushing. The patient was admitted to the ICU and received 1 L of normal saline and 2 units of O− packed RBCs. An indwelling urinary catheter was placed to assist in monitoring his volume status. After another large episode of hematemesis, Mr. M was intubated for airway protection. IV octreotide was begun, and the GI service was called to perform urgent endoscopy.
Have you crossed a diagnostic threshold for the leading hypothesis, variceal hemorrhage? Have you ruled out the active alternatives? Do other tests need to be done to exclude the alternative diagnoses?
The patient is having a large upper GI bleed and is clearly actively bleeding. Initial management is aimed at hemodynamic stabilization. The decision to place the patient in the ICU was based on his hemodynamic instability, active bleeding, and the need for close monitoring. Given the alcohol history, the volume of the bleed, and the lack of previous abdominal symptoms, bleeding from esophageal varices is highest on the differential diagnosis, and empiric therapy is begun with octreotide and antibiotics. Peptic ulcer disease is the most common cause of upper GI bleeding, and we do not yet know whether this patient has cirrhosis.
Alternative Diagnosis: Peptic Ulcer Disease
The details of peptic ulcer disease are discussed in Chapter 32, Unintentional Weight Loss. This section will only deal with hemorrhage from peptic ulcers.
Textbook Presentation
The classic presentation is a middle-aged person who has chronic dyspepsia; has used nonsteroidal anti-inflammatory drugs (NSAIDs), including aspirin, long-term; or who has Helicobacter pylori infection and an episode of hematemesis, melena, or both.
Disease Highlights
A. Most common cause of GI bleeds.
1. Upper GI bleeds are 4–8 times more common than lower GI bleeds.
2. Peptic ulcer disease accounts for at least 50% of upper GI bleeds.
B. Bleeding occurs when an ulcer erodes into a vessel in the stomach or duodenal wall.
C. About 50% of patients with bleeding or perforation have had no previous symptoms.
D. Causative factors are H pylori infection, long-term use of NSAIDs, or stress from critical illness.
E. Similar to diverticula, most cases are self-limited (about 80%).
Evidence-Based Diagnosis
A. Except in rare cases, all patients with GI bleeding in whom an ulcer is suspected undergo endoscopy. Endoscopy is useful from diagnostic, prognostic, and therapeutic standpoints.
B. Endoscopy has a 92% sensitivity for ulcers. Biopsy during endoscopy allows for the exclusion of malignancy and H pylori infection as a cause of the ulcer.
C. Endoscopy is also useful because it gives information about a patient’s risk of recurrent bleeding and thus enables discharge planning. Table 19-6 gives approximate rates for recurrent bleeding by endoscopic finding.
Table 19-6. Approximate rates for recurrent bleeding by endoscopic finding.
D. Other endoscopic findings associated with high risk are ulcer size > 2 cm and arterial bleeding.
Treatment
A. Hemodynamic stabilization
B. Endoscopy
1. Early endoscopy achieves hemostasis in > 94% of patients and decreases length of hospital stay.
2. For patients with a high-risk lesion, endoscopic intervention such as clipping, thermocoagulation, or sclerotherapy is warranted.
3. Repeat endoscopy is effective in the 15–20% of patients who have a recurrence of bleeding.
C. Medication
1. IV H2-blockers are no longer recommended for acutely bleeding ulcers.
2. Proton pump inhibitors
a. IV proton pump inhibitors should be given to patients admitted with suspected bleeding ulcers. Meta-analyses support the use of intermittent IV bolus or oral proton pump inhibitors; however, guidelines do not reflect this new data yet.
b. Patients found to be at high risk for rebleeding (Table 19-6) on endoscopy should continue this therapy for 72 hours.
c. Patients at low risk can be switched to oral proton pump inhibitors.
d. All patients who are discharged should be taking proton pump inhibitors (as well as H pylori therapy, if warranted) to ensure ulcer healing.
e. Patients taking aspirin for secondary prophylaxis can be restarted within 1 to 7 days of resolution of bleeding.
f. Meta-analysis supports administering erythromycin before endoscopy because it increases the diagnostic yield of the procedure and reduces the need to repeat endoscopy.
D. Patients who have rebleeding that cannot be controlled endoscopically can undergo either embolization or surgical therapy.
Alternative Diagnosis: Mallory-Weiss Tear
Textbook Presentation
Mallory-Weiss tear is typically seen in patients with vomiting of any cause in whom hematemesis develops acutely.
Disease Highlights
A. Mallory-Weiss tears are mucosal tears at the gastroesophageal junction.
B. It is a common misconception that Mallory-Weiss tears always follow retching. In fact, a history of retching preceding hematemesis is present in only about one-third of cases.
Evidence-Based Diagnosis
Diagnosis is routinely made on upper endoscopy.
Treatment
Mallory-Weiss tears seldom require specific treatment. Rebleeding is quite rare.
CASE RESOLUTION
Emergency endoscopy was performed in the ICU. Mr. M was found to have large esophageal and gastric varices. A clear bleeding source was found and treated with banding. Although there was no clinically significant rebleeding, other complications developed. He remained intubated for 5 days for presumed aspiration pneumonia, experienced alcohol withdrawal, and developed mild encephalopathy.
During the hospitalization, he was found to have Child-Pugh grade B cirrhosis. At the time of discharge, he was taking propranolol and lactulose. Follow-up in an outpatient alcohol program and the hepatology practice was scheduled. He did not come to any follow-up visits.
Mr. M’s emergent endoscopy was indicated by the severity of the bleeding. His bleeding was controlled with a combination of medical and endoscopic management. The complicated hospital course is not surprising given the comorbid conditions frequently present in patients with varices. Mr. M had advanced cirrhosis and alcohol dependence.
CHIEF COMPLAINT
PATIENT
Ms. S is a 35-year-old woman who comes to the outpatient clinic for an initial visit. She is well and is without complaints. On review of systems, she notes that she occasionally passes bright red blood per rectum. This has happened about 4 times over the past 5 years. It is never associated with pain. She sometimes sees the blood on the toilet paper and sometimes in the bowl.
At this point, what is the leading hypothesis, what are the active alternatives, and is there a must not miss diagnosis? Given this differential diagnosis, what tests should be ordered?
RANKING THE DIFFERENTIAL DIAGNOSIS
Ms. S has recurrent, lower GI bleeding that has occurred intermittently over a number of years without obvious negative health effects. This type of bleeding can be categorized as benign sounding anorectal bleeding. It is bleeding in a young patient without “red flags” for serious disease such as anemia, change in bowel habits, weight loss, or diarrhea. Between 10% and 20% of the population will have this type of bleeding. The goal is to diagnose these patients appropriately without missing occasional serious lesions and without subjecting excessive numbers of patients to unpleasant evaluation. The pivotal points, in this case, are the patient’s young age, the small volume of blood loss, and the absence of “red flags.”
The differential diagnosis includes hemorrhoidal bleeding and bleeding from anal fissures. Anal fissures are usually painful so hemorrhoids are the more likely diagnosis in this case. IBD, especially ulcerative colitis, could cause similar symptoms, but the intermittent nature of symptoms and characterization of the bleeding, (rare and small volume) make IBD less likely. We need to know more about the patient’s bowel habits. Diverticula and colonic angiodysplasia could account for the patient’s symptoms but would be very unusual in a patient this age. Colon or rectal cancer is also rare in this age group but should be considered. Table 19-7 lists the differential diagnosis.
Table 19-7. Diagnostic hypothesis for Ms. S.
On further history, Ms. S reports no recent change in bowel habits, no weight loss, and says she feels well. She does report that although the bleeding has never been associated with pain, it is sometimes associated with constipation. She has never used any treatment.
Is the clinical information sufficient to make a diagnosis? If not, what other information do you need?
Leading Hypothesis: Hemorrhoidal Bleeding
Textbook Presentation
Hemorrhoidal bleeding typically presents with rectal pain and bleeding. The pain is worst with bowel movements, straining, or sitting. Occasionally, hemorrhoids can present with painless bleeding.
Disease Highlights
A. Hemorrhoids are generally classified as external or internal.
1. External hemorrhoids
a. Occur below the dentate line.
b. Present either as painless bleeding or with engorged, painful, swollen perianal tissue; or with thrombosis. Thrombosed hemorrhoids are purple, extremely painful, and may bleed.
2. Internal hemorrhoids
a. Occur above the dentate line.
b. Symptoms can be a feeling of internal fullness, painless bleeding, or prolapse. Prolapse is usually painful and sometimes associated with bleeding.
3. Both internal and external hemorrhoids are most symptomatic when sitting, straining, and with constipation.
4. Bleeding occurs during defecation and commonly is seen on toilet paper when wiping.
A clinician should always verify a patient’s self-diagnosis of hemorrhoids. Many patients refer to all perianal symptoms as hemorrhoids.
Evidence-Based Diagnosis
A. Hemorrhoidal bleeding is diagnosed by direct observation.
1. This may be accomplished visually in patients with external hemorrhoids.
2. Patients with internal hemorrhoids require anoscopy to see hemorrhoids.
B. An important question is “When does benign sounding anorectal bleeding need a more extensive evaluation than an anal exam with or without anoscopy?”
1. One study looked at 201 patients whose review of symptoms revealed rectal bleeding.
a. 24% of these patients were found to have serious disease. The diseases were polyps (13%), colon cancer (6.5%), and IBD (4%).
b. Factors associated with risk of serious disease were age, short duration of bleeding, and blood mixed with stool.
c. No cancers were found in patients younger than 50.
d. 6 of the 37 patients who had a clear source of anorectal bleeding (fissures or hemorrhoids) also had polyps or cancer.
2. Another study found only 10 polyps among 314 patients under 40 with rectal bleeding compared with 27 polyps and 1 case of cancer among 256 patients between the ages of 40 and 50.
C. In general, if a young patient (under age 40) with rectal bleeding does not have a clear anorectal source or if the bleeding continues despite treatment of the anorectal source, a more complete evaluation (with colonoscopy) should be done. Patients over 40 should always be evaluated.
Although serious disease is rare among young people with rectal bleeding, it does occur.
Treatment
A. Most hemorrhoids and anal fissures can be treated conservatively with general recommendations for perianal well being.
1. Sitz baths to relax anal sphincter.
2. Analgesia with acetaminophen, topical creams, or short-term topical corticosteroids. A doughnut cushion is sometimes helpful for prolonged sitting.
3. Soften stool with increased fluid intake, a high-fiber diet, and docusate sodium or mineral oil.
4. Avoid anything that may lead to constipation.
5. Avoid prolonged sitting, especially on the toilet.
B. Internal hemorrhoids that prolapse or continue to bleed usually require surgical removal.
C. Thrombosed, irreducible internal hemorrhoids and thrombosed external hemorrhoids require rapid surgical treatment.
MAKING A DIAGNOSIS
Ms. S has a normal general physical exam. External anal exam and digital rectal exam are normal. Anoscopy reveals 1 large, nonbleeding internal hemorrhoid. A CBC is normal.
Have you crossed a diagnostic threshold for the leading hypothesis, hemorrhoidal bleeding? Have you ruled out the active alternatives? Do other tests need to be done to exclude the alternative diagnoses?
The patient has an internal hemorrhoid on exam. This is almost certainly, but not definitely, the cause of her bleeding. Because she is currently asymptomatic, it would be reasonable to postpone further work-up for now.
Alternative Diagnosis: Anal Fissures
Textbook Presentation
Patients typically have severe rectal pain with bowel movements and bright red blood on the toilet paper. On physical exam, a fissure can be found at the midline, posterior to the anal opening.
Disease Highlights
A. Anal fissures occur secondary to trauma to the mucosa of the anal canal, most commonly by hard stool.
B. Fissures usually present as acute onset, painful defecation, usually with bleeding.
C. Fissures can become chronic.
1. Pain causes anal sphincter spasm that, in turn, causes recurrent trauma.
2. Chronic fissures can be associated with sentinel piles (essentially a thickening of the skin at the end of a fissure).
D. Fissures are present at the midline.
1. Fissures are usually posterior in men and can be posterior or anterior in women.
2. Other diagnoses, such as Crohn disease or sexually transmitted diseases, should be considered when fissures are lateral to the anal opening.
Evidence-Based Diagnosis
A. Fissures are diagnosed by direct observation.
B. Physical exam is sometimes difficult since patients are often in pain.
Treatment
A. In most cases, general supportive recommendations outlined above for the treatment of hemorrhoids will bring relief of symptoms in days to weeks.
B. More chronic fissures often need therapy to relax the anal sphincter.
1. Topical nitrates and injected botulinum toxin are effective.
2. Surgical sphincterotomy is almost always effective but carries a small risk of permanent fecal incontinence.
CASE RESOLUTION
One year later Ms. S returns to the clinic with recurrent bleeding. Anoscopy revealed a bleeding internal hemorrhoid. Symptoms resolve with supportive care, but bleeding recurs 1 month later. Colonoscopy is performed and reveals only internal hemorrhoids. The patient declines definitive therapy and continues to experience rare episodes of hemorrhoidal bleeding.
The patient’s history of recurrent bleeding is quite common. Many patients with hemorrhoids will have occasional flares. The decision to perform colonoscopy was a difficult one. Although her young age and presence of an abnormality on anoscopy makes serious disease unlikely, evaluation of any patient with recurrent rectal bleeding is appropriate.
REVIEW OF OTHER IMPORTANT DISEASES
Occult GI Bleeding
Textbook Presentation
Occult GI bleeding presents in 1 of 2 ways: either in a patient with newly discovered iron deficiency anemia or in a patient with positive fecal occult blood tests.
Disease Highlights
A. Generally a disease of older patients; average age in most studies is the early 60s.
B. Upper GI lesions cause occult GI bleeding slightly more common than lower GI lesions.
C. The common causes of upper and lower GI bleeding account for most causes of occult GI bleeding.
D. Long-term aspirin, NSAID, or alcohol use is found in about 40% of patients with an upper GI tract lesion.
E. About 5% of patients have lesions of both the upper and lower GI tract.
Evidence-Based Diagnosis
A. All patients with occult GI bleeding need evaluation of the GI tract.
B. All patients with iron deficiency anemia need to have the cause of the iron deficiency identified (see Chapter 6, Anemia).
1. Iron deficiency is usually due to chronic blood loss. Rarely, it is due to poor iron intake or iron malabsorption.
2. Menstrual and GI blood loss are the most common sources.
3. All men, all women without menorrhagia, and all women over 50 (even those with menorrhagia) need to have an evaluation of the GI tract.
4. Women under age 40 with menorrhagia do not necessarily need further GI evaluation unless they have GI symptoms or a family history of early colon cancer.
5. Women between 40 and 50 years of age with menorrhagia need to be managed carefully. They should be asked about minimal GI symptoms (celiac disease causes iron deficiency through malabsorption and the symptoms can be easily attributed to irritable bowel syndrome). There should be a low threshold for recommending colonoscopy in this subset of patients.
Always determine the source of blood loss in occult GI bleeding and iron deficiency anemia.
C. Evaluation of the GI tract in patients with occult GI bleeding should be done as follows:
1. Colonoscopy alone is sufficient in patients with only a positive fecal occult blood test and no iron deficiency.
2. Evaluation of entire GI tract is recommended by most experts for those patients with iron deficiency anemia or iron deficiency anemia and positive fecal occult blood testing.
3. Video capsule endoscopy should be used in patients in whom EGD and colonoscopy are inconclusive.
Small Bowel Bleeding
Textbook Presentation
Patients with small bowel bleeding may present with either occult or overt bleeding. Diagnosis of a small bowel source is made during small bowel imaging or endoscopy after negative evaluation of the colon and upper GI tract.
Disease Highlights
A. 5–10% of patients with GI bleeding have a negative evaluation of the upper and lower GI tract.
1. With current endoscopic and radiologic techniques, approximately 75% of these patients will be found to have a small bowel source of bleeding.
2. The remainder of cases have lesions of the upper or lower GI tract missed on initial EGD and colonoscopy.
B. Sources of bleeding
1. Angiodysplasia is the most common source of bleeding in the small bowel.
2. Crohn disease, Dieulafoy lesions, malignancy (accounting for about 10% of small bowel bleeding), and Meckel diverticula are other common causes.
C. Obscure GI bleeding, which used to refer to GI bleeding with negative upper and lower endoscopy, now only refers to GI bleeding with normal upper and lower endoscopy as well as small bowel evaluation by advanced endoscopic or radiographic procedures.
Evidence-Based Diagnosis
A. A directed history may provide clues to the source of small bowel or obscure GI bleeding. Ask about the use of medications that can cause mucosal damage (eg, NSAIDs, bisphosphonates) as well as a history of diseases that predispose patients to GI bleeding (HIV, neurofibromatosis).
B. Repeat endoscopy, looking for lesions that were missed on the initial evaluation, is often the initial step in the diagnosis.
C. If repeated upper and lower endoscopies are negative, or if there is a high clinical suspicion for a small lesion, enteroscopy and video capsule endoscopy are used to image the small bowel.
D. There are multiple types of enteroscopy, including
1. Push enteroscopy, in which a long endoscope (often a colonoscope) is passed orally.
a. Visualization of 40–60 cm of jejunum is common.
b. Diagnostic yields of 40–75% have been reported.
2. Deep small bowel enteroscopy (double balloon enteroscopy, single balloon enteroscopy, and spiral enteroscopy) can achieve visualization of the entire small bowel.
3. Intraoperative endoscopy.
E. Meckel diverticulum scan uses a nuclear tracer that binds to parietal cells.
1. Sensitivity for detecting a Meckel diverticulum is between 75% and 100%.
2. Diagnosis only really considered when obscure bleeding occurs in a patient younger than 30.
Treatment
The treatment of obscure bleeding varies by the cause of bleeding.
REFERENCES
Barkun AN, Bardou M, Kuipers EJ; International Consensus Upper Gastrointestinal Bleeding Conference Group. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med. 2010;152(2):101–13.
Barkun AN, Bardou M, Martel M et al. Prokinetics in acute upper GI bleeding: a meta-analysis. Gastrointest Endosc. 2010;72:1138–45.
Bull-henry K, Al-kawas F. Evaluation of occult gastrointestinal bleeding. Am Fam Physician. 2013;87(6):430–6.
Cuellar RE, Gavaler JS, Alexander JA et al. Gastrointestinal tract hemorrhage. The value of a nasogastric aspirate. Arch Intern Med. 1990;150(7):1381–4.
Garcia-Tsao G, Bosch J. Review article: Management of varices and variceal hemorrhage in cirrhosis. N Engl J Med. 2010;362:823–32.
Gerson LB, Fidler JF, Cave DR, Leighton JA. ACG clinical guideline: Diagnosis and management of small bowel bleeding. Am J Gastroenterol 2015;110:1265–87.
Gralnek IM, Dumonceau JM, Kuiper EJ et al. Diagnosis and management of non-variceal upper gastrointestinal hemorrhage: European Society of Gastrointestinal Endoscopy Guideline. Endoscopy. 2015;47(10):a1–46.
Helfand M, Marton KI, Zimmer-Gembeck MJ, Sox HC Jr. History of visible rectal bleeding in a primary care population. Initial assessment and 10-year follow-up. JAMA. 1997;277(1):44–8.
Huang ES, Karsan S, Kanwal F, Singh I, Makhani M, Spiegel BM. Impact of nasogastric lavage on outcomes in acute GI bleeding. Gastrointest Endosc. 2011;74(5):971–80.
Jacobs D. Clinical practice: Hemorrhoids. N Engl J Med. 2014;371:941–51.
Laine L. Clinical practice: upper gastrointestinal bleeding due to a peptic ulcer. N Engl J Med. 2016;374:2367–76.
Lewis JD, Shih CE, Blecker D. Endoscopy for hematochezia in patients under 50 years of age. Dig Dis Sci. 2001;46(12):2660–5.
Madhotra R, Mulcahy HE, Willner I, Reuben A. Prediction of esophageal varices in patients with cirrhosis. J Clin Gastroenterol. 2002;34(1):81–5.
McGuire HH Jr. Bleeding colonic diverticula. A reappraisal of natural history and management. Ann Surg. 1994;220(5):653–6.
Ohmann C, Thon K, Stoltzing H, Yang Q, Lorenz W. Upper gastrointestinal tract bleeding: assessing the diagnostic contributions of the history and clinical findings. Med Decis Making. 1986;6(4):208–15.
Olds GD, Cooper GS, Chak A, Sivak MV Jr, Chitale AA, Wong RC. The yield of bleeding scans in acute lower gastrointestinal hemorrhage. J Clin Gastroenterol. 2005;39(4):273–7.
Pickhardt PJ, Choi JR, Hwang I et al. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Engl J Med. 2003;349(23):2191–200.
Raju GS, Gerson L, Das A, Lewis B; American Gastroenterological Association. American Gastroenterological Association (AGA) Institute technical review on obscure gastrointestinal bleeding. Gastroenterology. 2007;133(5):1697–717.
Sachar H, Vaidya K, Laine L. Intermittent vs continuous proton pump inhibitor therapy for high risk bleeding ulcers. JAMA Intern Med. 2014;174(11):1755–62.
Sharara AI, Rockey DC. Gastroesophageal variceal hemorrhage. N Engl J Med. 2001;345(9):669–81.
Srygley FD, Gerardo CJ, Tran T, Fisher DA. Does this patient have a severe upper gastrointestinal bleed? JAMA. 2012;307(10):1072–9.
Stanley AJ, Laine L, Dalton HR et al. Comparison of risk scoring systems for patients presenting with upper gastrointestinal bleeding: international multicentre prospective study. BMJ. 2017;356:i6432.
Strate LL. Lower GI bleeding: epidemiology and diagnosis. Gastroenterol Clin North Am. 2005;34(4):643–64.
Strate LL, Gralnek I. ACG clinical guideline: management of patients with acute lower gastrointestinal bleeding. Am J Gastroenterol. 2016;111:459–74.
Villanueva C, Colomo A, Bosch A et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368(1):11–21.
Witting MD, Magder L, Heins AE et al. ED predictors of upper gastrointestinal tract bleeding in patients without hematemesis. Am J Emerg Med. 2006;24(3):280–5.