Page 26 - Gastrointestinal Bleeding (Xuất huyết tiêu hóa)

300 PART III Symptoms, Signs, and Biopsychosocial Issues

2 units of packed RBCs or a decrease in the hematocrit value of recommended (see Chapter 129). Most patients, however, espe-
20% or more) or recurrent bleeding after 24 hours of stability cially if older than 50 years of age, will also require colonoscopy,
(defined as the need for additional transfusions, a further decrease at least electively, to evaluate the remainder of the colon.
in the hematocrit value of at least 20%, or readmission for LGI
bleeding within 1 week of discharge). Predictive factors include Flexible Sigmoidoscopy
tachycardia, hypotension, syncope, a nontender abdomen, wit-
nessed rectal bleeding on presentation, aspirin use, and more than Flexible sigmoidoscopy can evaluate the rectum and left side of
2 comorbid illnesses. 244,245 These risk factors are used in a prog- the colon for a bleeding site and can be performed without a
nostic scoring system that identifies patients at the highest risk for standard colonoscopy bowel preparation. Although not adequate
severe LGI bleeding, who account for 19% of patients with LGI for evaluation of the anal canal, flexible sigmoidoscopy alone
bleeding and may benefit most from urgent colonoscopy. will result in a diagnosis in approximately 9% of cases. 250 If the
A single-institution case series of 94 patients admitted for LGI distal colon can be adequately cleansed with enemas, an urgent
bleeding 246 found that 39% of all cases of LGI bleeding requir- flexible sigmoidoscopy can be useful for patients suspected of
ing hospitalization were severe, as defined by the passage of red having a solitary rectal ulcer, UC, radiation proctitis, postpolyp-
blood after the patient had left the emergency department and ectomy bleeding (in the rectosigmoid), or internal hemorrhoids
associated hypotension or tachycardia or the need for a transfu- (see Chapters 41, 116, 119, 126, 128, and 129). Therapeutic
sion of more than 2 units of packed RBCs during hospitalization. hemostasis can be provided with injection therapy, hemoclip
Independent risk factors for severe LGI bleeding were an initial placement, band ligation, or MPEC. Monopolar electrocautery
hematocrit value of 35% or lower, abnormal vital signs (a systolic (e.g., argon plasma coagulation, snare polypectomy, hot biopsy
blood pressure <100 mm Hg or a heart rate >100/min) on admis- forceps) should not be used if a bowel preparation has not been
sion, and gross blood on initial rectal examination. administered to avoid the risk of ignited flammable colonic gas
Artificial neural networks have also been used to develop pre- (see Chapter 17).
diction models for severe LGI bleeding, 247,248 but from a clinical
point of view, the large number of variables that have to be entered Radionuclide Imaging
into a computer program for analysis limit their widespread use.
Radionuclide imaging involves injecting a radiolabeled sub-
Mortality stance into the patient’s bloodstream and performing serial
scintigraphy to detect focal collections of radiolabeled mate-
A large U.S. database study of 227,000 patients with a discharge rial (see earlier). This technique has been reported to detect
diagnosis of LGI bleeding reported an overall mortality rate of bleeding at a rate as low as 0.04 mL/min, with an overall
48
3.9% in 2008. 240 Multivariate analysis found that independent positive diagnostic rate of approximately 45% and an accuracy
predictors of in-hospital mortality are age older than 70 years, rate of 78% for localizing the true bleeding site. 238 The disad-
intestinal ischemia, at least 2 comorbid illnesses, onset of bleed- vantages of radionuclide imaging are that delayed scans may
ing after hospitalization for an unrelated condition, coagulopa- be misleading, and determining the specific cause of bleeding
thy, hypovolemia, transfusion of packed RBCs, and male gender. often depends on endoscopy or surgery. False-positive results
Colorectal polyps and hemorrhoids were associated with a lower are most likely to occur when transit of luminal blood is
mortality risk. The low risk of death from LGI bleeding identi- rapid, so that radiolabeled blood is detected in the colon even
fied in this study is consistent with data from smaller series such though it originated in the UGI tract. Radionuclide imaging
as those from Kaiser San Diego (2.4%) and the University of may be helpful in cases of obscure GI bleeding (see later) or
California, San Francisco (3.2%). 235,246 The Kaiser study also prior to angiography to help localize a lesion, particularly if
found an increased risk of death with in-hospital LGI bleeding. an early scan (e.g., 30 minutes to 4 hours after injection of the
radiolabeled material) is positive for RBC extravasation.
Diagnostic and Therapeutic Approach
Patients with hematochezia should undergo the same care- Angiography
ful history taking, physical examination, and laboratory testing Angiography is most likely to detect a site of bleeding when the
45
described earlier for the general approach to the patient with rate of arterial bleeding is at least 0.5 mL/min. The diagnostic
acute GI bleeding (see Table 20.1). The history should focus yield depends on patient selection, the timing of the procedure,
specifically on identifying sources of LGI bleeding. Diverticular and the skill of the angiographer, with positive results in 12%
bleeding should be suspected in patients with painless, severe, to 69% of cases. An advantage of angiography is that emboliza-
acute hematochezia and a history of diverticulosis, although isch- tion can be performed to control some bleeding lesions. Major
emic colitis may also be painless. 249 complications, however, occur in 3% of cases and include bowel
Patients should be medically resuscitated. Because LGI bleed- ischemia, hematoma formation, femoral artery thrombosis, con-
ing is generally less severe than UGI bleeding, blood transfu- trast dye reactions, acute kidney injury, and transient ischemic
47
sions may not be required. Most patients should undergo initial attacks. Other disadvantages of angiography are the absence of
evaluation with colonoscopy after bowel preparation, although active bleeding in most patients at the time of angiography, inabil-
in selected cases anoscopy or flexible sigmoidoscopy without ity to detect nonbleeding SRH (NBVV, clot, or spot), expense of
any bowel cleansing or after an enema may be performed. Other the test, and inability to determine the specific lesion responsible
diagnostic tests, including radionuclide bleeding scans or angiog- for bleeding in many cases. 236,237 A small retrospective case series
raphy, may be used in selected cases or when colonoscopy fails to of 11 patients with colonic bleeding who underwent angiographic
detect a source of bleeding. embolization reported that the bleeding ceased in 10, mesen-
teric ischemia developed in 7, and 6 died. 251 Another study of 65
Anoscopy patients with acute LGI bleeding who did not undergo colonos-
copy as a first diagnostic step found that diagnostic angiography
Anoscopy can be useful for patients in whom bleeding internal provided little additional clinical information because the bleed-
hemorrhoids or other anorectal disorders (e.g., fissures, fistulas, ing stopped spontaneously in most patients. Moreover, angiog-
proctitis) are suspected from the medical history. For internal raphy did not help guide subsequent surgery and was associated
hemorrhoids, immediate treatment with rubber band ligation is with a complication rate of 11%. 252